/•

BIOLOGICAL BULLETIN

OF THE

flTmrine Biological laboratory

WOODS HOLE, MASS.

Editorial Staff

E. G. CONKLIN — Princeton University.

GEORGE T. MOORE — The Missouri Botanic Garden,

T. H. MORGAN — Columbia University.

W. M. WHEELER — Harvard University.

E. B. WILSON — Columbia University.

Managing ]£fcitor

FRANK R. LILLIE — -The University of Chicago.

VOLUME XL.

WOODS HOLE, MASS.
JANUARY TO JUNE, 1921

PRESS OF

THE NEW ERA PRINTING COMPANY
LANCASTER, PA.

CONTENTS OF VOLUME XL

No. i. JANUARY, 1921

PAGE

LILLIE, FRANK R. Studies of Fertilization. VIII, On the
Measure of Specificity in Fertilization Between Two
Associated Species of the Sea- Urchin Genus Strongylo-
centrotus I

LILLIE, FRANK R. Studies of Fertilization. IX, On the
Question of Superposition of Fertilization on Partheno-
genesis in Strongylocentrotus purpuratus 23

HYMAN, LIBBIE H. The Metabolic Gradients of Vertebrate

Embryos 32

No. 2. FEBRUARY, 1921 *

UHLENHUTH, EDWARD. Observations on the Distribution and
Habits of the Blind Texan Cave Salamander, Typhlomolge
Rathbuni 73

NEWMAN, H. H. On the Development of the Spontaneously

Partheno genetic Eggs of Asterina (Patiria) Miniata .... 105

NEWMAN, H. H. On the Occurrence of Paired Madreporic
Pores and Pore-Canals in the Advanced Bipennaria
Larvce of Asterina (Patiria) Miniata together with a
Discussion of the Significance of Similar Structures in
Other Echinoderm Larvce 118

No. 3. MARCH, 1921

HUXLEY, JULIAN S. Differences in Viability in Different
Types of Regenerates from Dissociated Sponges, with
a Note on the Entry of Somatic Cells by Spermatozoa ... 127

HYDE, I. H. A Micro-Electrode and Unicellular Stimulation 130

STARR, ISAAC, JR. Effect of Variations in Oxygen Tension on

the Toxicity of Sodium Chloride Isotonic to Sea Water ... 1 34

LOEB, LEO. Transplantation and Individuality 143

iii

IV CONTENTS

No. 4. APRIL, 1921

HUBBS, CARL L. The Ecology and Life-History of Amphigo-
nopterus Aurora and of Other Viviparous Perches of
California . 181

ALLEN,. WILLIAM RAY. Studies of the Biology of Freshwater

Mussels 210

No. 5. MAY, 1921

HARRIS, J. A., AND REED, H. S. Inter-Periodic Correlation

in the Analysis of Growth 243

SCHRADER, FRANZ. The Chromosomes of Pseudococcus

Nipa 259

JUDAY, CHANCY. Observations on the Larvce of Corethra

Punctipennis Say 271

No. 6. JUNE, 1921

THOMAS, L. J. Morphology and Orientation of the Otocysts of

Gonionemus 287

'BRUES, CHARLES T., AND GLASER, RUDOLF W. A Symbiotic
Fungus Occurring in the Fat-Body of Pulvinaria in-
numerabilis Rath 299

RICHARDS, A., AND THOMPSON, JAMES T. The Migration of

the Primary Sex-Cells of Fundulus heteroclitus 325

HONDA, H. Spermatogenesis of Aphids ; The Fate of the

Smaller Secondary Spermatocyte 349

Vol. XL. January, 1921 No. i.

BIOLOGICAL BULLETIN

AUTHOR S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 7

STUDIES OF FERTILIZATION.

VIII. ON THE MEASURE OF SPECIFICITY IN FERTILIZATION
BETWEEN Two ASSOCIATED SPECIES OF THE SEA-URCHIN

GENUS STRONGYLOCENTROTUS.1

FRANK R. LILLIE,
UNIVERSITY OF CHICAGO.

CONTENTS.

I. Introduction i

II. Comparison of the Gametes 3

III. Cross-fertilization.

1. General 3

2. Variability of different individual cross-fertilizations 7

(a) Purpuratus eggs fertilized by franciscanus sperm 7

(&) Franciscanus eggs fertilized by purpuratus sperm 8

3. The effect of sperm concentration 8

(a) Purpuratus § X franciscanus <$ 8

(fc) Franciscanus £ X purpuratus $ p

4. The measure of specificity 10

5. Fertilization with mixtures of eggs and sperm 13

IV. Cross-agglutination 14

V. General 20

I. INTRODUCTION.

In previous studies on fertilization the writer has maintained
that the substance from eggs that agglutinates spermatozoa of the
same species is necessary for fertilization; more specifically that
the activating effect of the spermatozoon in fertilization is pri-
marily on this substance, which then sets the other processes in
operation. The agglutinating substance was accordingly called

1 The writer wishes to express his indebtedness to J. Nelson Gowanlock for
his devoted and careful work as assistant in these experiments and in those
of study IX. T

2 FRANK R. LILLIE.

fertilisin. If there is such a connection between the capacity of
the eggs for agglutinating spermatozoa and for being fertilized,
then, in addition to other consequences which have been followed
out by Just (1919), Moore (1916, 1917) and the writer (1914,
1919), the fertilization reaction and sperm agglutination should
show a similar degree of specificity. This has been shown to be
the case in wide crosses, as between Arbacia and Nereis by the
writer, and between Arbacia and Echinarachnius by Just. But
the matter has not been tested with two species of a single genus
where the test might be expected to be crucial. If there were no
connection between agglutination of spermatozoa and fertilization
of the egg, the phenomena need not exhibit similar specificities;
but if the phenomena were found to be similarly specific we would
have a new and strong argument for the postulated connection.

The problem of specificity in fertilization is at present obscure,
and in any case data presenting the nature and degree of spe-
cificity could not fail to be useful. It is rather extraordinary
that no attempt to secure even roughly quantitative results on this
problem in animals has hitherto been recorded.

In the neighborhood of Pacific Grove, California, Strongylo-
centrotus purpuratus and 5\ franciscanus are found in protected
situations along the rocky shores. S. purpuratus occurs a short
distance below the high-water mark and extends to an unde-
termined depth, certainly several feet, below the low-water mark.
S. franciscanus rarely occurs above the low-water mark and cer-
tainly extends into deeper water than purpuratus. At low water
one thus finds purpuratus in isolated tide pools and sometimes
entirely exposed, whereas franciscanus is very rarely found in
such situations. The two s.pecies are commonly, though by no
means always, intermingled just below the low water mark. In
one collecting station purpuratus was so abundant in this zone
as to form a veritable pavement on the floor and a covering on
the sides of the partially open rock-pool ; S. franciscanus, very
conspicuous by its larger size, longer spines and different color-
ing, was interspersed among the individuals of the other species.

At the time of my visit, January and February, 1920, both
species were ripe simultaneously at this station and elsewhere;
but whereas all individuals of purpuratus had practically only

STUDIES OF FERTILIZATION. 3

ripe gametes wherever found, franciscanns at this station and
elsewhere, with one exception, had relatively undeveloped gonads,
which, however, always contained ripe gametes, often in con-
siderable abundance, which fertilized readily. At the exceptional
station, well within Monterey Bay, all individuals of both species
had perfectly developed gonads. The waiter wishes to express
his appreciation of the hospitality of the Hopkins Marine Sta-
tion during this investigation, and his thanks to the Director, Dr.
W. K. Fisher, for his aid in securing material and in many other
essential ways.

II. COMPARISON OF GAMETES.

6\ franciscanns is much larger than 5\ purpuratus, commonly
almost twice the diameter, and its much larger spines cause the
impression of an even greater discrepancy in size. It is an in-
teresting fact that the gametes of the larger species are almost
correspondingly larger than those of the smaller, both in the case
of the ova and also the spermatozoa. The eggs of franciscanns
are no to 114 /JL in diameter, the eggs of purpuratus from 75 to*
79 fj.. The head of the spermatozoon of franciscanus is about
7ju, long by 2/x broad at the base, that of purpuratus is about 4 by
2 p. The jelly surrounding the franciscanns egg is about 30^ in
thickness and relatively firm and resistant, that surrounding the
egg of pnrpuratns is about 15/1 in thickness, relatively soft and
easily lost.

III. CROSS-FERTILIZATION.
i. General.

Loeb's recorded observations contain the only information in
the literature on the subject of the cross-fertilization of these two
species. They are to the following effect: (i) " If we mix eggs
of franciscanns and purpuratus in sea-water and add the sperm of
purpuratus the eggs of purpuratus will be fertilized more quickly
than the eggs of franciscanus; and the reverse is true if the sperm
of franciscanus is added to a mixture of both eggs in sea-water '
(p. 273, Am. Nat., 1915). (2) "The sperm of purpuratus
shows no trace of cluster formation with the egg-sea water of
franciscanus, and yet the eggs of franciscanus are readily fertil-

4 FRANK R. LILLIE.

ized with the sperm of purpuratus." (Loeb, 1914, p. 136.)

(3) In his "Artificial Parthenogenesis and Fertilization" (The
University of Chicago Press, 1913), p. 293, Loeb gives figures
of plutei from purpuratus eggs fertilized by franciscanus sperm.
These records of the occurrence of both reciprocal fertilizations
give, however, no measure of the degree of specificity; in fact,
there is only a bare hint that the straight fertilizations occur more
readily than the crosses.

If one wishes to gain a correct idea of the degree of specificity
in fertilization, it is necessary to control the variable factors very
carefully. Assuming that only perfectly ripe gametes are used,
as was the case in all the recorded experiments, the principal pre-
cautions to be observed are the following: (i) The eggs should
be washed in at least two changes of sea-water to get rid of tissue
secretions, blood or detritus ; the eggs used should be uniform
in this respect, and similar quantities should be used in com-
parable experiments. (2) The sperm should be sufficiently abun-
dant so that measurable quantities of the dry sperm free from
any fluid or foreign particles may be used as a basis for cal-
culating the sperm concentration in any experiment. (3) The
range of individual variability with reference to cross-fertiliza-
tion should be understood; it is at least very considerable.

(4) The method of mixing the sperm with the eggs should be
as uniform as possible, or very considerable differences in per-
centage results may occur from this cause alone. (5) It should
go without saying that the utmost precautions against contamina-
tion must be observed : abundance of sterilized pipettes and glass-
ware, washing of specimens in drinking water to destroy adherent
spermatozoa, sterilization of hands and instruments after handling
any male, etc. (6) In spite of all precautions there will remain
a certain degree of residual variability that shows clearly that all
the conditions of fertilization are not yet understood.

Sperm Concentrations. — The basic sperm suspension from
which fertilizations were usually made is one drop (o.i c.c.) dry
sperm thoroughly mixed in 5 c.c. sea-water; one drop (0.07 c.c.)
of this suspension added to eggs in 100 c.c. sea-water will be arbi-
trarily selected as it nit sperm concentration. Any measured in-
semination may be expressed as unity or as a fraction or multiple

STUDIES OF FERTILIZATION. 5

thereof. The various inseminations are all so designated. Unit
sperm concentration would usually be called a "light insemina-
tion " and ten unit concentration would usually be regarded as
fairly " heavy." The absolute value of " unity " would be one
part of the dry sperm in about 70,000 parts of sea-water.

There is not much difference in the relative fertility of the
crosses between the two species. In the records given it appears
that franciscanus eggs may give a small percentage of cross
fertilization at lower sperm concentrations than purpuratus eggs ;
but the purpuratus spermatozoon is much smaller than the fran-
ciscanus spermatozoon, so that it is probable that sperm suspen-
sions of purpuratus rated as of the same concentration as those of
franciscanus really contain a very much greater number of sperm-
atozoa, which would tend to explain the lower concentration for
minimum results in cross fertilization of franciscanus eggs. On
the other hand, the highest percentage of fertilizations recorded
in a cross (Table 4) concerns purpuratus eggs ; this may be due
to the much greater number of experiments with this cross than
with the reciprocal.

The cross-fertilized purpuratus eggs appear to be more viable
than the cross-fertilized franciscanus eggs ; the latter do not in
my experience develop to the pluteus stage, while the former do
readily enough. The blastulae even of the cross-fertilized fran-
ciscanus eggs ofter appear abnormal, and the gastrulae very
commonly so. It is doubtful to what factor to ascribe this differ-
ence, whether to the large size of the cytoplasmic mass of the
franciscanus egg and small size of the purpuratus spermatozoon,
or to behavior of the chromosomes in the reciprocal crosses.
Material is on hand for investigating the latter possibility.

Another difference noted in the two reciprocal crosses is that
the cross-fertilized purpuratus eggs commonly form as fine mem-
branes as the straight fertilized ones, whereas in a much higher
percentage of franciscanus eggs the membranes are either " tight,"
in the sense that they do not stand out so far from the surface
of the egg, as in the straight fertilization, or lacking entirely.
Now the tight membrane and absence of membrane in fertilized
eggs are signs of poor condition or low vitality. This difference,

6 FRANK R. LILLIE.

which is far from being absolute, would appear to indicate that as
a rule the franciscanus egg is understimulated by the purpuratus
spermatozoon, and this may aid to explain their lesser viability as
compared with the reciprocal cross.

When high sperm concentrations are used (5 units and above)
the jelly of the crossed franciscanus eggs becomes permanently
packed with purpuratus spermatozoa, thus producing the appear-
ance of halos varying in intensity with the concentration of the
sperm. Such halos do not appear in the reciprocal cross, nor yet
in either of the straight fertilizations. The difference in the
physical characters of the jelly in the two kinds of egg does not
explain this difference, for the franciscanus sperm does not form
comparable halos in franciscanus eggs. It would appear to be an
effect of some secretion contained in the franciscanus jelly on the
purpuratus sperm; but the franciscanus egg-water has no appar-
ent effect on the purpuratus sperm.

We shall consider first the variability of different individual
combinations in the two crosses separately, classified again by
the sperm concentration ; secondly, we .shall consider the effect
of various sperm concentrations on the percentages of fertiliza-
tion in the eggs of one female ; third, we shall tabulate the com-
plete experiments in which both crosses and both controls were
run simultaneously. The first table will give the measure of
variability of different individual cross combinations. The
second will give the effects of sperm concentrations for given
combinations. The third will give the measure of specificity,
which cannot be evaluated without the other two.

The percentages of fertilization may be measured either by the
percentage of membranes formed, or by the percentage of eggs
that actually segment. The latter determination is much more
accurate, for membrane formation is apt to be defective in the
cross-fertilized eggs, or even absent, more especially where fran-
ciscanus eggs are used. In some cases, both determinations were
made, and the difference gives approximately the percentage of
membraneless eggs that segment. Eggs that form membranes
sometimes fail to segment, though this is relatively rare ; it ap-
pears especially in the lower sperm concentrations.

STUDIES OF FERTILIZATION.

2. Variability of Different Individual Cross Fertilisations.

(a) Purpuratus Eggs Fertilised by Franciscanus Sperm. — For
the purpose of bringing out the individual variability of cross
combinations we shall consider first the fertilizations made with
sperm of 10 units concentration (Table I) ; temperature approxi-
mately 15° C. Each entry stands for a separate female.

TABLE I.

VARIATION OF FERTILITY OF PURPURATUS EGGS FERTILIZED WITH 10 UNITS

FRANCISCANUS SPERM.

No.

Percentage of
Membranes.

Percentage of
Cleavage.

Remarks.

I

12%

One franciscanus male

2 ....
4

5-.. •
6
7. .. .
8

1.6%

13' %

4 %
17 %
19 %
o

4%
16%

29%
39%
72%

84%

One franciscanus male

i %

10

i %

> One franciscanus male

12

6.7%

'

One franciscanus male

2O.I%

One franciscanus male

The variations of fertility in this cross, from o to 84 per cent,
of cleavage, is exceeding striking and shows a great range to
exist on the female side (cf. especially Nos. 2-7 where one fran-
ciscanus male was used). There is a probable considerable
variation also on the male side as appears when the group 2-7
fertilized with one franciscanus male is compared with the group
8-1 1 fertilized with another franciscanus male. But no system-
atic attempt was made to evaluate this side, as the franciscanus
material was relatively rare and difficult to obtain. It should also
be noted that such high percentages as Nos. 6 and 7 were never
approximated in any other combination (43 in all recorded and
many others tried).

To correct the impression that this cross-fertilization is very
easy, which might readily be given by the exceptional percentages
in Nos. 2-7, the following record of the fertilization of the eggs

8 FRANK R. LILLIE.

of five purpuratus with the sperm of one franciscanus at 13.2
units will be useful: percentages segmented o.i per cent., 0.3 per
cent., O.6 per cent., 1.6 per cent., 6 per cent.

It should be realized that sperm concentrations of 10 and more
units are very much higher than would ordinarily be used in
straight fertilizations, where I unit concentration is adequate to
fertilize every egg.

(&) Franciscanus Eggs Fertilised by purpuratus Sperm. — The
determinations for this cross are not nearly so numerous as for
the reciprocal on account of the greater rarity of the material in
its best condition. The eggs appear to be not quite so variable
in their cross-fertilization capacity as those of purpuratus, but
with a lower concentration of sperm they commonly show a
higher percentage of cleavage after cross-fertilization than cross-
fertilized purpuratus eggs; on the other hand, I have never
secured so high a percentage with higher sperm concentrations
as in exceptional lots of purpuratus eggs. Moreover, the cross-
fertilized franciscanus eggs appear to be less viable than the
cross-fertilized purpuratus eggs. These matters will be dealt
with beyond. The data given in Table IV. throw some light on
the question of variability.

3. The Effect of Sperm Concentration.

In the case of both species a higher concentration of sperm is
required for any degree of cross fertilization than for practically
perfect straight fertilization (see Table IV.). Beyond such
minimum concentration, the result of increased sperm concen-
tration is to increase the percentage of eggs fertilized up to a cer-
tain point; however, such increase is by no means proportional
to the sperm concentration and sometimes it is very strikingly
absent.

(a) 5. purpuratus 5 X S. franciscanus <$. — Though the higher
percentages of fertilization were never secured with sperm con-
centration below about 6 units, yet it must be said that the factor
of individual variability of combinations is much more important
than that of sperm concentration. An isolated illustration will

STUDIES OF FERTILIZATION.

TABLE II.

EFFECT OF SPERM CONCENTRATION, PURPURATUS 9. X FRANCISCANUS

Strength of Sperm.

Percentage of
Membranes.

Percentage of Seg-
mented Eggs.

Remarks.

I unit

2.5%

I+% \

2 units

2-5%

I+% /

Eggs of one female

8 '

<o.S% 1

Eggs of one female

32 '

<o.s% /

10 '

12% 1

20 '

20% \

Eggs of one female

40 '

25% J

prove this better perhaps than the figures given in Table II : On
February 6 I tested the cross-fertilization of a given lot of pur-
puratus eggs with franciscanus sperm up to a concentration of 75
units with hardly any fertilization resulting at any concentration.
The purpuratus eggs thus resistant to franciscanus sperm at all
concentrations fertilized perfectly with purpuratus sperm (added
later), and the franciscanus sperm used fertilized franciscanus
eggs perfectly.

Another experiment in which the eggs of five purpuratus
females were fertilized in each case with franciscanus sperm of
6.6 and 13.2 units will bring out once more the greater effective-
ness of the individual variations than the sperm concentration :

TABLE III.

EFFECT OF SPERM CONCENTRATION ON EGGS OF FIVE FEMALES.
S. purpuratus 5x5". franciscanus $.

Percentage Segmented with
13.2 Units Sperm.

6

1.6
0.6

O.I

0.3

There is a much greater difference between females I and 4
for instance than between the two lots of eggs of any one female.

(b) S. franciscanus $ X «S\ purpuratus £. — In the case of this
cross again the individual variation outruns the effects of sperm
concentration. Thus four samples of the eggs of one female
fertilized with i, 2, 4 and 8 units of the sperm of one purpuratus

Percentage Segmented with
6.6 Units Sperm.

9. i

4.2

0 2

0.4

9 1

0.9

94.

O."?

9°;

o..-?

IO FRANK R. LILLIE.

gave 0.3 per cent., 0.6 per cent., 1.3 per cent., and 4 per cent,
cleavage respectively. Two samples of the eggs of another
female fertilized with i and 2 units of the sperm of another pur-
puratus gave in each case 5 per cent, cleavage. In a third case
two samples of the eggs of one female fertilized with 8 and 32
units of purpuratus sperm gave about 6 and 5 per cent, mem-
branes respectively; and in a fourth case three samples of the
eggs of one female fertilized with 10, 20 and 40 units purpuratus
sperm gave 30 per cent., 5o per cent., and 65 per cent, cleavage.

4. The Measure of Specificity.

From the preceding data it will be seen that the measure of
specificity varies greatly with the individuals used, and this is
a fact that must be born in mind in considering the following
data. But however much variation there may be in the crosses,
they contrast forcibly with the straight fertilizations whether
they are considered individually or collectively.

The experiments tabulated below (Table IV.) were set up as
follows : the usual precautions were observed for sterilization of
animals, instruments, hands, glassware ; each dish had a separate
sterilized pipette used for it alone. The eggs of one female of
each species were always washed at least twice, and an unfertil-
ized control of each lot was kept, which never showed any mem-
branes or segmented eggs. One hundred c.c. of sea water was
placed in the bowls and about 1.5 c.c. of eggs was transferred to
each, and the bowls then covered with glass plates. When the
eggs were all ready for fertilization, the sperm was prepared in
another room, and brought in and added as rapidly as possible ;
one lot of each kind of sperm thus always served for all the
fertilizations of a given experiment. My assistant removed the
cover, stirred in the sperm as added and replaced the cover before
proceeding to the next bowl. In each experiment in the table
only one male and one female of each species was used; straight
fertilizations with the same male and female thus serve for
control of the condition of the material used, and also for esti-
mates as to the degree of specificity.

STUDIES OF FERTILIZATION.

II

TABLE IV.

THE MEASURE OF SPECIFICITY IN CROSSES OF STRONGYLOCENTROTUS FRANCIS-

CANUS AND S. PURPURATUS.

M. = percentage of membranes formed.
S. = percentage of eggs segmented.

Sperm

F$XI

V.

F9 :

<Pcf.

P9 X

Pd".

P9 X

Fcf.

Concen-
trations,

M.

S.

M.

S.

M.

S.

M.

S.

Exp. i:

a

O.O5

73-3%

o

0

17 %

0

o

b

0.2=;

95 %

o

o

36.5%

o

0

c. . . .

\J

o.s

QQ %

o

<0.2%

56.5%

0

0

Exp. 2:

*-" J

y y / 1/

a

0.25

100 %

4-6%

33 %

0

b

O.^

100 %

4-6%

65 %

0

c. .

w<o
I.

100 %

4-6%

75 %

<O.I%

Exp. 3:

*•"-"-' / U

a

I

85-90 %

80%

3%

5%

IOO %

100%

i.5%

1.25%

b

2

85-00 %

80%

3%

5+%

IOO %

100%

i-5%

1-25%

«-• j X / U

/ \s

\j / \J

J I / \J

Exp. 4:

a

8

IOO %

5%

IOO %

<o.5%

* W /^

J / 1/

<o-5%

b

32

5%

Exp. 5:

•

a

10

100%

Mem-

30%

100%

Mem-

12%

b

20

100%

branes

50%

100%

branes

20%

c

40

100%

noted

65%

100%

noted

25%

as

as

tight

good

The above data show that the eggs of each species straight
fertilize well at a sperm concentration which has no effect in cross
fertilization. The fertilization is thus strikingly specific to an
extent that is not realized when higher sperm concentrations are
employed. If we take the lowest sperm concentration in each
case in which any cross fertilization occurs, viz., 0.25 in the case
of franciscanus eggs (Exp. 2a) and I in the case of purpuratus
eggs (Exp. 2c) we can give a mathematical expression to the
relative " ease " of straight and cross fertilization in a given lot
of eggs. Thus in the case of franciscanus where 100 per cent,
of the eggs fertilized with 0.25 franciscanus sperm we could say
that cross fertilization was 20 times more " difficult " than straight
fertilization. Or if we use Exp. ic for comparison, cross-fert.il-

12 FRANK R. LILLIE.

ization would be reckoned 500 times more difficult than straight
fertilization in the case of franciscanus eggs. In the case of
purpuratus where 75 per cent, fertilized with I unit purpuratus
sperm and <o.i per cent, fertilized with I unit franciscanus
sperm we could say similarly that cross-fertilization was at least
750 times more " difficult " than straight fertilization. It is of
course doubtful how much validity such a form of expression
possesses, but it is probably sufficient to emphasize the enormous
difference that actually exists, which a reader who has not worked
with the material might otherwise fail to appreciate.

Attempts were made to overcome the very strong specificity;
increase in concentration of the sperm has only a very limited
effect as already noted. The influence of increase of temperature
and of alkalinity of the sea water was also investigated in both
species with strikingly negative results: Temperature — (i) Pur-
puratus. Samples of eggs of one female were fertilized with
sperm of one male franciscanus (10 units strength) at 15° C, 17°
C, 24° C. and 27° C. with the following percentages of eggs
segmenting: 15°, 6.3 per cent.; 17°, 10.7 per cent.; 20.5°, 3.9 per
cent.; 24°, 3.5 per cent.; 27°, 2.2 per cent. (2) Franciscanus.
Samples of eggs of one female were fertilized with sperm of one
male purpuratus (8 units strength) at 14° C., 17.5° C, 20.5° C.,
27° C. with the following percentage of eggs segmenting: 14°,
3 per cent.; 17.5°, 6 per cent.; 20.5°, n per cent.; 27°, <i
per cent.

Alkali. — (i) Purpuratus. Sample of eggs were fertilized with
franciscanus sperm (10 units strength) in (a) 50 c.c. sea water
+ 0.4 c.c. N/io KOH, (&) +0.6 c.c. N/io KOH, (c) -f 0.8 c.c.
N/io KOH, (d) + i c.c. N/io KOH, (*) + 1.2 c.c. N/io KOH.
In none of these was there any increase above the control, and in
the last two decided decreases. (2) Franciscanus. A similar
series fertilized with purpuratus sperm of 4 units concentration
using NaOH instead of KOH; no fertilization occurred in any
of the hyperalkaline solutions ; the control showed 0.4 per cent.
fertilization.

STUDIES OF FERTILIZATION. 13

5. Fertilisation with Mixtures of Eggs and Sperm.

Loeb has stated that if mixed eggs of the two species are
fertilized separately with the two kinds of sperm the eggs of the
same species as the sperm form their membranes more quickly
than those of the other species. This would appear to be a
promising method of measuring specificity provided that normally
the membrane reaction is equally rapid in the two species. There
is, however, a considerable difference in this respect : the eggs of
purpuratus form membranes about 30 seconds after straight in-
semination, whereas the larger eggs of franciscanus require about
90 seconds (at about 17° C.)- Thus when mixed eggs of the
two species are fertilized with purpuratus sperm, the purpuratus
membranes naturally form first ; when the franciscanus sperm is
used the question is difficult to answer decisively, because the pur-
puratus eggs that fertilize are usually so few that it is difficult to
find them in the time available. My impression is that a small
percentage forms membranes in about the usual time, but others
more slowly so that after the franciscanus membranes are all
formed some purpuratus membranes still arise. In this respect,
as in others, there is great variability in the crosses.

The question involved would be more simply answered by
timing the formation of membranes in straight and cross fertiliza-
tions. The latter determinations were not made systematically,
but in those cases in which I have records of membrane forma-
tion of crosses taken some minutes apart, I find the percentages
higher in later than in earlier determinations. Thus it would
appear that while in a straight insemination membranes appear to
form in most of the eggs close together in time, in cross insemina-
tions membrane formation is spread over a longer period of time.

The above observations suggested the fertilization of mixed
eggs not only with the sperm of each species separately but also
with mixed sperm of the two species in order to determine
whether there was any antagonism between the sperm of the two
species. A rather elaborate experiment of this kind showed con-
clusively that no antagonism exists, such as is found sometimes
between sperm suspensions of different phyla. Thus a mixture

14 FRANK R. LILLIE.

of equal parts of both kinds of sperm (2 drops dry sperm to 5 c.c.
of sea water in each case) was made one day at 11.58 A.M. and
was used to fertilize mixed eggs at 12.02 P.M., at 12.46, at 1.49,
at 2.43 P.M. and at 10.54 A.M. the next day. In the first fertil-
ization practically all eggs of both kinds fertilized ; in the second
fertilization about 55 per cent, of the franciscanus eggs fertilized
and 75 per cent, of the purpuratus; percentages were not esti-
mated in the third and fourth fertilizations, but there was little
if any falling off. In the last fertilization made about 23 hours
after mixing, using the same eggs and sperm there was practically
no fertilization. Controls consisting of the same mixed eggs
fertilized with the same sperm samples used in the experiment,
but unmixed, were run, showing approximately the same rate of
falling off as the mixed sperm, during the first day. There is
thus no evidence of any antagonism of sperm. On the second
day the unmixed franciscanus sperm fertilized about 15-20 per
cent, of the franciscanus eggs in the mixture, but the mixed
franciscanus sperm fertilized practically none. The purpuratus
eggs did not fertilize either with unmixed or mixed purpuratus
sperm. There is thus an indication that the franciscanus sperm
loses fertilizing power more rapidly in presence of purpuratus
sperm than when separate.

IV. CROSS- AGGLUTINATION.

In 1913 I described the phenomenon of agglutination of sperm
suspensions of Arbacia by the egg-water of the same species ;
the phenomenon is a reversible one, the duration of which depends
on the concentration of the egg-water. With stronger solutions
the. agglutinated masses are larger and last longer; with the
lowest effective dilution the masses are microscopic in size and
disperse in 5 or 6 seconds. It is possible by using a minimum
quantity of water to a large bulk of eggs to secure an egg-water
that will give the minimum reaction when diluted 1,200 times.
The phenomenon is associated with an intense activation of the
spermatozoa, due to a distinct substance, and is evidently caused
by a change in the heads of the spermatozoa that renders them

STUDIES OF FERTILIZATION. 15

i

sticky (adhesive), as an important factor at least. It is obvious
that the reaction will detect only such changes in the individual
spermatozoa as will cause them to adhere to one another; and
that below this threshold the individual spermatozoa are prob-
ably affected to some extent ; this consideration is important for
the theoretical deductions.

I found also that the egg-water of the Annelid Nereis simi-
larly agglutinates Nereis spermatozoa. But the agglutinating
substances of Arbacia and of Nereis are without effect on the
heterologous sperm. The reaction is thus specific between these
forms. The egg-water of Arbacia, however, contains a sub-
stance toxic to the spermatozoa of Nereis and causing a coagula-
tion of Nereis sperm suspensions that is rather deceptive at first
sight. It is possible to remove this substance by treatment of the
egg-water with Nereis sperm and leave the full complement of
specific agglutinating substance, which can then be shown to be
without visible action on Nereis sperm. The Nereis egg-water
contains nothing that acts on Arbacia, sperm.

Just (''19) subsequently showed that similar specific relations
of agglutination obtained between Arbacia and Echinarachnius:
" Echinarachnius egg-water activates but does not agglutinate
Arbacia sperm. Arbacia egg-water agglutinates Echinarachnius
sperm. This is a ' hetero-agglutination ' by a substance in Ar-
bacia egg-water separate from Arbacia fertilizin because it may
be removed from the egg-water by dilution, by repeatedly wash-
ing the eggs and by precipitating it with Echinarachnius sperm.
It is found in Arbacia blood. The microscopic appearance of
this toxic heteroagglutination of Echinarachnius sperm is dif-
ferent from that of the iso-agglutination."

It was of considerable interest to discover how specific this
agglutinating reaction really is. If the reaction indicates a sub-
stance contained in the eggs that is of fundamental significance
in fertilization, as I believe, then it should be highly specific, like
the fertilization reaction itself ; and so it turns out, as the follow-
ing data show.

The egg-waters of both species have strong agglutinating ac-
tion, each on the sperm of its own species, which is quite similar

1 6 FRANK R. LILLIE.

in all respects to agglutination in Arbacia previously described.
Such egg-waters vary in strength according to the condition of
the eggs used and the relative amounts of eggs to the sea-water.
With perfectly fresh eggs added to about three times their own
bulk of sea-water, the egg-water of either species may attain a
strength of 640 agglutinating units, i.e., such an egg-water may
be diluted 640 times and still cause a minimum agglutination re-
action in sperm suspensions of its own species.

The egg-waters of S. purpuratus are more variable in strength
than those of 5\ franciscanus due to the fact that the jelly layer,
which contains the larger part of the detectable agglutinating
substance, is relatively very soft in purpuratus eggs and is rapidly
lost by washing the eggs. Thus washed eggs of purpuratus
rarely give as high an egg-water test as washed eggs of
franciscanus.1

The agglutinating reaction is so prompt and unmistakable that
it was really a great surprise to find that egg-waters of maximum
strength may be entirely devoid of agglutinating action on the
sperm of the other species, even undiluted.

The egg water of S. franciscanus never has the slightest ag-
glutinating action on the sperm of 5\ purpuratus. The absence
of reaction in this case was so invariable that it seems unneces-
sary to cite instances.

On the other hand, the egg-water of the majority of individuals
of purpuratus may cause an apparent agglutination of the
sperm of franciscanus; so that we may have the appearance of a
lack of agreement between the reciprocals in this case just as we
found between Arbacia and Nereis, and between Arbacia and
Echinarachnius (Just). The egg-water from certain purpuratus

i 1 have confirmed Loeb's observation that eggs of purpuratus from which
the jelly has been removed by acid do not produce sufficient agglutinating sub-
stance to agglutinate the spermatozoa. Such eggs are nevertheless capable of
fertilization as Loeb has stated. The eggs of franciscanus, however, after
removal of the jelly by acid continue to produce amounts of agglutinating
substance sufficient to agglutinate its own spermatozoa. In this respect fran-
ciscanus agrees with Arbacia (Lillie, 1915). I should regard the result in the
case of purpuratus as showing that an amount of fertilizin insufficient for
sperm agglutination is yet adequate for fertilization; and would reject Loeb's
ready interpretation that the fertilizin is not necessary for fertilization.

STUDIES OF FERTILIZATION. 1 7

females has, however, not the slightest effect on franciscanus
sperm, however concentrated it may be. (See Table VII., No.
6.) This is not at all unusual. The hetero-agglutinating action
is sporadic, while the iso-agglutinating action is constant. On
the assumption that the iso-agglutinating action is caused by a
single substance we cannot have this substance sometimes agglu-
tinating the sperm of franciscanus and sometimes not, unless we
assume that the variability is in the sperm of different individuals
of franciscanus. This cannot be the case, for the same sperm
suspension of franciscanus may show the apparent agglutination
with the egg-water of one individual of purpuratus and fail to
show it with another equally strong in isoagglutination. I have
also tried a non-heteroactive egg-water of purpuratus on the
sperm of different individuals of franciscanus and found it nega-
tive consistently. It must therefore follow that the egg-waters
of certain indivaduals of purpuratus contain a separate hetero-
active substance.

This substance is apparently not a constituent of the normal
blood of purpuratus, for the blood of purpuratus seems to be as
indifferent a medium for the sperm of franciscanus as for the
sperm of purpuratus, and the reverse is also true. In this respect
it differs from the substance of Arbacia egg-water active on
Nereis and Echinarachnins.

It should also be noted that the apparent hetero-agglutination
differs in two other important respects from either iso-agglutina-
tion, (i) in the appearance of the reaction, (2) in the relation
of duration of the reaction to dilution :

In iso-agglutination round solid masses of agglutinated sperma-
tozoa form in a few seconds ; these are preceded by strand-
formations which rapidly contract, and the strands must be pre-
ceded by cloud formations which, however, condense so rapidly
to strands as to make positive observation of these exceedingly
difficult. This is true both with stronger and weaker solutions,
the difference consisting in the scale of the phenomena and the
rate of reversal. The action of purpuratus egg-water on fran-
ciscanus sperm never proceeds beyond the stage of dense clouds ;
it does not therefore resemble the action of more dilute egg-

18

FRANK R. LILLIE.

waters on own sperm, for the cloud phase is never passed in the
hetero-agglutination.

In the second place, the duration of the iso-agglutination fol-
lows a very definite rule in relation to concentration of the
agglutinating substance which Richards and Woodward (1915)
have formulated by saying that the efficiency of the agglutinin
varies with the square root of the concentration. It is certain
that the duration of the agglutination is always greatly increased
with doubling of concentration, and in the lower concentrations
of I to 4 units nearly doubled.

Table V. gives one set of results in which the egg- water of six
females of purpuratus was tested for duration of the reaction on
franciscanus sperm at four dilutions.

TABLE V.

DURATION OF HETEROAGGLUTINATION. EGG-WATERS OF Six PURPURATUS FEMALES

ON FRANCISCANUS SPERM AT VARIOUS DILUTIONS.

Dilution.

i/i.

1/8.

1/24.

1/80.

I

10 sees.

IS '
6 '

9 '

IS '

20 '

8 sees.

10 '

18 '
ii '
16 '

12 '

neg.
faint
10 sees,
neg.
8 sees.
8 "

neg.

14

?

neg.

t i

4 I

2

•I . .

A .

5. .

6..

It will be seen that these figures do not follow any definite rule
with reference to duration of reaction and concentration, and
thus contrast in the most striking way with the iso-agglutina-
tion phenomena. This may be because the action at its optimum
is so slight on the individual spermatozoa of franciscanus as to
preclude any close agglutination. The indicator (i.e., the fran-
ciscanus sperm suspension) may thus appear to show the same
reaction at different dilutions of the egg-water.

On the assumption that we are concerned with a single hetero-
active substance distinct from the true agglutinin, it should be
possible to remove this substance by treating the egg-water con-
taining it with franciscanus sperm without disturbing the iso-

STUDIES OF FERTILIZATION.

agglutinating power of the egg-water in question. This experi-
ment was tried twice with the general result that the action on
both kinds of sperm was gradually destroyed by franciscanus
sperm, but the heteroactive substance disappeared more rapidly.
Thus no clear cut result was obtained.

The determined facts might conceivably find an explanation on
the assumption that the action of purpuratus fertilizin on fran-
ciscanus sperm varies with a state of aggregation of the pur-
puratus fertilizin. I do not, however, find such an explanation
satisfactory because under experimental conditions, apparently
identical, certain samples of purpuratus fertilizin show this effect
and another does not, and there is no comparable variation for
the iso-agglutination.

The eggs used in the cross-agglutination experiment cited
previously (Table V.) were fertilized uniformly with francis-
canus sperm (10 units) with the following results:

TABLE VI.

COMPARISON OF CROSS-AGGLUTINATION AND FERTILIZATION.

Purpuratus 9.

Cross Agglutination
at 1/8 Dilutions.

Fertilized with Franciscanus Sperm.

% Membranes.

% Cleavage.

I

8 sees.
10
18
ii
16

12

19%
1-7%
17%
13%
1.6%
4 %

84%
16%
72%
29%
4%
39%

2

-J .

4.

c

6

A second experiment in which the same eggs of six purpuratus
were used for cross-agglutination and cross-fertilization gave the
following results (Table VII.).

TABLE VII.

A SECOND EXPERIMENT ON CROSS-AGGLUTINATION AND FERTILIZATION.

Cross Agglutination. Fertilized with/rana'scanus Sperm.

1 strong 0.3 % segmented.

2 " 7-4%

3 " o

4 " 0.2%

5 " 14-9%

6 . absent . o

it
it

2O FRANK R. LILLIE.

It is thus impossible to establish any relationship between cross
agglutination and cross fertilization ; i.e., to predict from the
cross-agglutination of franciscanus sperm by purpuratus egg-
water what the success of the cross-fertilization of these eggs
might be. This result is to be expected from the conclusion that
the apparent cross agglutination is caused by a substance distinct
from the purpuratus fertilizin, that is, by a substance not directly
concerned in fertilization.

GENERAL.

The outstanding result of the investigation is that the reaction
of the egg to the spermatozoon and the reaction of the sperma-
tozoa, by agglutination, to egg secretions are both highly specific.
Fertilization and agglutination run in the same direction, and this
gives strong new additional evidence that there is a specific con-
nection between fertilization and the capacity for agglutination
of the spermatozoa. It is in fact hardly conceivable that such a
reciprocal specificity between egg and spermatozoon should be
without significance.

Let me now deal with the apparent difficulty that agglutina-
tion (the reaction of the sperm to egg-secretions) appears to be
more specific than fertilization (the reaction of the egg to the
spermatozoon). As I have already pointed out in the present
paper, the agglutination reaction will detect only such changes in
the individual spermatozoa as will cause them to adhere to one
another; any change below this threshold will fail to be dis-
tinguished. On the other hand, the fertilization reaction is to be
detected in the individual egg. A higher degree of detectible
specificity may therefore be expected for the agglutination reac-
tion than for the fertilization reaction.

Secondly, the adhesion of spermatozoa to one another is no
part of the process of fertilization ; indeed, such adhesion could
not occur in the sperm concentrations ordinarily used in fertiliza-
tion, or occurring in nature, owing to the wide dispersion of the
spermatozoa. It is the change in the individual spermatozoon
and the postulated reciprocal change in the fertilizin that is

STUDIES OF FERTILIZATION. 21

significant, not the adhesion of spermatozoa to one another that
results in sufficiently concentrated sperm suspensions.

One cannot therefore argue from the absence of the agglutina-
tion reaction in any given case that there is no connection between
the modification of the spermatozoon evidenced by agglutination
and fertilization. This is admittedly a serious limitation of the
method for analysis of fertilization, but the limitation does not
weaken the force of the positive results.

My previous interpretation of the phenomena of fertilization
was that the substance that agglutinates the spermatozoa is the
same as that which activates the egg. I therefore named this
substance fertilizin, for its presence is essential for fertilization.
The fertilization reaction was thus conceived to be a combination
of a sperm component ("sperm receptors") with the fertilizin
of the egg, and a consequent activation of the fertilizin which
then reacts with egg components ("egg receptors") initiating
the developmental processes.

This conception furnishes a schema within which the complex
events and immediate consequences of fertilization can be readily
comprised. The original arguments (Lillie, 1914) for the neces-
sity of fertilizin in the fertilization reaction were as follows :
(i) The production of fertilizin by eggs ceases at fertilization;
correspondingly the eggs are "immune" to spermatozoa. (2)
The production of fertilizin by eggs ceases after membrane
formation by butyric acid; correspondingly these eggs are not
fertilizable. (3) If the fertilizin content of eggs be reduced by
repeated washings the fertilization capacity decreases correspond-
ingly. (4) The production of fertilizin by eggs of the sea-urchin
does not begin until after maturation; correspondingly the fully
grown ovocytes with intact germinal vesicle are immune to
spermatozoa, or, if spermatozoa enter such eggs, they are entirely
without effect.

These results have been confirmed and extended by Just
(1919) and by Moore (1916, 1917). The latter author made a
special study of the butyric acid reaction with special reference
to the question of superposition of fertilization ; while Just has
dealt with all of these points, especially in the case of Echin-

22 FRANK R. LILLIE.

arachnws, and has shown a constant relationship of a very
definite sort between fertilizin content of the eggs as shown by
the agglutination reaction and their capacity for fertilization.

The measurements of specificity contained in the present paper
are interpretable in the same way. One cannot regard the spe-
cificity in fertilization and in sperm agglutination as unrelated
phenomena. The change in the spermatozoon produced by the
fertilizin of its own species must in some way be connected with
the relative ease with which the spermatozoon fertilizes the egg
of its own species. The fertilizin theory gives a reasonable view
of this relationship. The variable percentages of cross-fertiliza-
tion would show that a degree of reaction which would not
suffice for agglutionation of the spermatozoa may yet be suffi-
cient under favorable circumstances for activation of fertilizin.

LITERATURE.
Just, E. E.

'19 The Fertilization Reaction in Echinarachnius parma. I. Cortical Re-
sponse of the Egg to Insemination. II. The Role of Fertilizin in
Straight and Cross Fertilization. III. The Nature of the Activation of
the Egg by Butyric Acid. BIOL. BULL., Vol. 36, pp. 1-53.
Lillie, Frank R.

'13 Studies of Fertilization. V. The Behavior of Spermatozoa of Nereis
and Arbacia with Special Reference to Egg-extractives. Journ. Exp.
Zool., Vol. 14.

'14 Studies of Fertilization. VI. The Mechanism of Fertilization in Ar-
bacia. Journ. Exp. Zool., Vol. 16.

'19 Problems of Fertilization. The University of Chicago Press.
Loeb, J.

'14 Cluster Formation in Spermatozoa caused by Specific Substances from
Eggs. Journ. Exp. Zool., Vol. 16.

'15 On the Nature of the Conditions which Determine or Prevent the
Entrance of the Spermatozoon Into the Egg. Am. Nat., Vol.49, pp.
257-285.
Moore, C. R.

'16 On the Superposition of Fertilization on Parthenogenesis. BIOL. BULL.,
Vol. 31.

'17 On the Capacity for Fertilization after Initiation of Development.

BIOL. BULL., Vol. 32.
Richards, A., and Woodward, A. E.

'15 Note on the Effect of X-Radiation on Fertilizin. BIOL. BULL., Vol. 28.

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY J

STUDIES OF FERTILIZATION.

IX. ON THE QUESTION OF SUPERPOSITION OF FERTILIZATION ON
PARTHENOGENESIS IN STRONGYLOCENTROTUS PURPURATUS.

FRANK R. LILLIE,
UNIVERSITY OF CHICAGO.

The statement of Loeb (1913, p. 234; 1915, p. 260-261) that
eggs of Strongylocentrotus purpuratus which have formed mem-
branes as a result of butyric acid treatment can be fertilized with
sperm if the membranes are destroyed by shaking raises a diffi-
cult question in the problem of fertilization. It is known that
eggs in which membranes have been formed by spermatozoa are
incapable of refertilization, even if the membranes are destroyed
by shaking immediately after their formation. We have every
reason to believe that the membrane-forming reaction by butyric
acid is the same as by fertilization. The after effects should
therefore be the same.

In an examination of this problem in the case of Arbacia, C. R.
Moore (1916) showed that this is the case, viz: that, given a
full membrane reaction by butyric acid, the eggs became in-
capable of fertilization, even if the membranes are removed. It
is, however, possible to superimpose fertilization on an incom-
plete reaction caused by butyric acid to a variable extent which
is roughly proportional to the degree of the original reaction.
Just (1919) investigated the same problem in the case of
Echinarachnlus parva, and determined that eggs which have
formed full membranes after butyric acid treatment do not re-
spond to subsequent insemination whether the membranes are
removed or not.

The writer took advantage of the opportunity afforded by a
stay at the Hopkins Marine Station in Pacific Grove, California,
in January and February, 1920, to repeat Loeb's experiments on
the same species that he used. The results obtained diverged

23

24 FRANK R. LILLIE.

considerably from Loeb's account and they are therefore re-
corded for the purpose of clearing up the apparent flat contra-
diction between the results of Loeb on Strongylocentrotus pur-
puratus and of Moore and Just on other forms.

The conditions of activation by butyric acid should first be re-
called. It is convenient to proceed from a standard concentra-
tion of butyric acid, and accordingly the strength used by Loeb
was adopted for all the experiments, viz : 2.8 c.c. N/io butyric
acid -f- 5° c-c- sea-water. Membranes are not formed in butyric
acid, but only after transfer of the eggs to sea-water. The
optimum time of exposure to butyric acid for membrane forma-
tion in sea-water is a function of temperature ; a series of de-
terminations showed the following approximate optimum ex-
posures: 15° C, 110-120 seconds; 17°, 90 seconds; 20°, 80
seconds ; 24.5°, 40 seconds ; 30°, 18 seconds. It is therefore very
important that the temperature of any given experiment should
be known ; the temperature of the experiments recorded here was
always close to 16° C. The optimum time of exposure is also
rather narrowly limited at any temperature. Thus in one experi-
ment where 105 seconds was optimum, 90 seconds was too short
and 120 seconds too long for optimum results.

The concentration of the butyric acid to which the eggs are
exposed is another vital factor; therefore in adding the eggs to
the butyric acid the same amount of sea-water should be carried
over with them each time; in my experiments this was 2.5 c.c.'
The membrane reaction after transfer to sea-water depends on
removal of the action of the butyric acid ; it is therefore im-
portant not to carry over too much butyric acid in making this
transfer; the transfers were therefore made uniformly to TOO c.c.
sea-water and not more than 2.5 c.c. of the butyric acid solution
was carried over. The eggs were not fertilized in this sea-water,
but from 2 to 5 drops were transferred to 10 c.c. of fresh sea-
water in a Syracuse watch crystal in which the inseminations
were made ; thus the butyric acid was reduced to a negligible
quantity in the insemination.

Considerable variations in temperature or of conditions of
transfer to or from the butyric acid will produce considerable

STUDIES OF FERTILIZATION. 25

variations in results. But if an adequate series of controls is
kept it is possible to check these up.

The following controls were kept: (i) Some untreated eggs.
(2) Eggs left in the butyric acid. (3) Eggs left in the sea-water
to which transfer was made from butyric acid (unfertilized).
(4) When eggs were shaken for superimposed insemination,
some were left unfertilized. (5) In some cases samples from
control 3 were fertilized without shaking. With these controls
one can follow back the history of any given result and determine
the percentage and kind of membranes originally formed, the
extent to which they were destroyed by shaking, and the capacity
for fertilization without shaking.

The eggs used for any given experiment were selected usually
from several lots representing different females ; those eggs that
gave the best reaction to insemination being selected.

As Loeb and others have noted, the eggs that have received the
butyric acid treatment do not segment without subsequent treat-
ment, but invariably cytolyze, or die from other causes, sooner
or later. This happened invariably in my controls (3) and not
a single segmenting egg was recorded from such controls.

The membranes when first formed are soft, and are easily
destroyed by shaking within two minutes (approximately) after
transfer to sea-water from the butyric acid; but they soon
harden and thereafter it is difficult to destroy them by shaking.
Moreover, the membranes do not form instantaneously after
transfer to sea-water, but require about the same time (approxi-
mately 30 seconds at 15° C.) as for their formation by insemina-
tion. The time for effective results from shaking is therefore
cut down accordingly.

The exposure to butyric acid thus acts like insemination only
in the sense that it prepares the egg for membrane formation,
but the reaction does not begin until the butyric acid is removed ;
and, after that, the time required for membrane formation is the
same as for membrane formation of control eggs by insemina-
tion. So long as the eggs are in butyric acid, therefore, they
presumably maintain a f ertilizable condition, and it requires about
30 seconds after transfer for this fertilizable condition to become
lost bv occurrence of the full membrane reaction.

26 FRANK R. LILLIE.

If the exposure to butyric acid is not sufficient to produce the
membrane reaction after transfer to sea water the eggs are mostly
fertilizable as though they had not been exposed to the acid. If
the reaction after transfer to sea-water is incomplete, so that
membranes are not fully formed, some eggs retain a certain
amount of capacity for fertilization after the membranes are
removed so that they may segment; they do not, however, pro-
ceed far with the developmental process under such circum-
stances, rarely to the gastrula stage. Moreover, the eggs of any
lot are variable, within rather narrow limits, with respect to the
degree of reaction after any given exposure. It would therefore
hardly be expected that all eggs of any given lot would give the
optimum membrane reaction after any given exposure, and one
would expect a certain minimal capacity for segmentation after
superimposed insemination with the best possible lot of eggs.

Even in Loeb's experiments the outstanding fact is not the one
that he emphasizes, viz: that a certain percentage of eggs may
segment after superimposed fertilization when the membranes are
removed, but that most of the eggs do not do so. Loeb reports
as high as 20 per cent, segmentation after such a superimposed
insemination ; 80 per cent, of the eggs even in this case had be-
come unfertilizable, and this is the significant result. I have
never obtained such a high percentage of segmenting eggs in any
of my experiments after the optimum exposure, and my con-
clusion is that in this particular experiment of Loeb's the eggs
were on the whole under-exposed to butyric acid. I am strength-
ened in this opinion by the fact that he reports that the develop-
ing eggs formed normal larvce (1915, p. 261), a result that is
never obtained, in my experience, after an optimum exposure.
The experiment in question is reported as though only one ex-
posure time were used, in which case it would be a matter of
good fortune to hit the exact exposure time. In my experiments
at least three exposure times were always used by transferring
lots of eggs to sea water at 15 second intervals before, at, and
after the estimated time of optimum exposure ; then treating the
entire series.

Reference should be made to the tabulated results of the prin-
cipal experiments (pp. 29-31) for the following discussion:

STUDIES OF FERTILIZATION. 2J

It will be seen from experiments 2 and 7 that at 15-16° C. the
unshaken eggs have practically full fertilization capacity up to a
3O-second exposure to butyric acid and that the transition to the
unfertilizable condition is very sudden — thus between 30 and 45
seconds (Exp. 7) and between 30 and 60 seconds (Exp. 2). This
transition coincides with the onset of membrane formation and
all eggs that have formed membranes fail to fertilize. This is
well known from previous experiments, but it is not generally
recognized that many eggs that fail to form membranes also
become unfertilizable. Thus in experiment 2c after 60 seconds
exposure to butyric acid 15 per cent, of the eggs failed to form
membranes, but only 4.9 per cent, of all eggs segmented after
insemination. In experiment 7 this comes out even more clearly :
30 per cent of the eggs failed to form membranes in 7& (after
45 seconds exposure) and only 0.2 per cent, of all eggs segmented
after insemination.

If the eggs be shaken within about 2.V-2. minutes after transfer
to sea-water the soft membranes are destroyed ; there is however
a rapid transition from this soft condition of the membrane to a
hard condition in which it is difficult to destroy the membranes
by shaking. Rather complete destruction of membranes in the
soft condition is accompanied by extensive agglutination of the
eggs, some, however, remaining free. Agglutination of the eggs
is indeed a fair index of the degree to which the membranes have
been destroyed.

When the membranes are thus destroyed there is a slight in-
crease in the percentage of eggs that segment after insemination
as compared with unshaken eggs (cf . Exp. 2, d, e; Exp. 3 ; Exp.
8). The percentage is in any case exceedingly small, and the
impressive fact is that in all cases over 95 per cent, of the eggs,
indeed generally over 99 per cent, of the eggs, are incapable of
segmenting after insemination. Although the percentage is so
small it must, I think, be admitted as possible that there is a small
percentage of eggs that have formed membranes that still possess
the capacity of segmenting after insemination. In other words, I
think that this small percentage of segmenting eggs is not neces-
sarily to be referred to eggs that failed in the membrane reaction,

28 FRANK R. LILLIE.

for such eggs should fertilize without shaking, and this they fail
to do as the unshaken inseminated control shows. The eggs that
segment do not form a new membrane and they fail to develop
far ; in the loose type of cleavage and irregularity of development
they exhibit a weak condition.

The rather significant fact will be noted in the tables that the
small percentages of segmenting eggs involved diminish in each
experiment with increased time of exposure : thus in Exp. 2 : 0.6
per cent, after 1^2 minutes exposure, < o.i per cent, after 2
minutes exposure; Exp. 3: 1.3 per cent, after il/2 minutes, 0.15
per cent, after 1^4 minutes, and 0.25 per cent, after 2 minutes
(the slight increase here is not significant, and is probably due to
sampling) ; Exp. 8: 0.3 per cent., 0.15 per cent, and o after i%,
il/2 and 134 minutes respectively; Exp. 9: 4 per cent., 1.5 per
cent., 0.3 per cent, after 2, 2^4 and 2l/2 minutes respectively. (In
Exp. 9 it will be noted in the protocol that membrane formation
was not very complete.) There is a tendency in each experiment
towards a vanishing point. This can mean only that each egg
tends towards an unfertilizable state under the conditions of the
experiment and that the small residuum shown in the experiments
is due to the statistical conditions of individual variation.

In each experiment the eggs were very heavily inseminated,
i o to 20 times as much sperm being used as required for normal
fertilization. The spermatozoa did not, however, penetrate into
the experimental eggs except in rare cases, as shown by sections.
In those cases in which they were observed in sections to have
entered, apparently usually at points of injury to the egg-surface,
the spermatozoa remained unaltered in the cytoplasm. The yet
rarer cases in which a reaction resulting in cleavage was set up in
the egg were not observed in the sections, for the painstaking
search necessary to find them in such a mass of material did not
seem worth while.

My results agree in principle with those of Moore and Just
and are in disaccord with those of Loeb. It will be noted from
the tables that the optimum time of exposure to butyric acid varies
considerably even at the same temperature with different lots of

STUDIES OF FERTILIZATION. 2Q

eggs. Failure to take this into account may make considerable
difference in the results of superimposed insemination. The fact
that Loeb secured 20 per cent, of vigorous larvae in the principal
experiment that he cites suggests an under exposure, for the eggs
inseminated after the optimum exposure to butyric acid, that
segment, never develop normally.

Conclusion. — The membrane reaction after butyric acid treat-
ment is the same as after insemination ; this is shown by simi-
larity of the membranes formed in the two cases, and by the fact
that the rate of formation is the same. The result of rendering
the egg insusceptible to spermatozoa shows a possible slight
difference only, and this receives an obvious explanation on the
assumption that there is a variable tendency towards incomplete-
ness of reaction after butyric acid which in rare cases leaves some
of the reacting substances of the egg (fertilizin) free for sperm
action.

EXP. 2. TEMPERATURE ESTIMATED 15-16° C.

Exposure to Butyric.

Fertilized Without
Shaking.

Shaken After Mem-
brane Formation.

Fertilized after Shaking.

a. 15 sees

QQ+%

15 min.

78% segmented

b. 30

08 %

1C "

80% segmented

c. 60 "

4.0%

15 "

°-7% segmented

d. oo "

o

i \Q min.

0.6% segmented

e. 120

o

21

<o.i% segmented

/.ISO "

o

e. 240 "

o

h. 360 "

o

Membranes formed as follows after butyric : a, o ; b, o ; c. 85 per cent. ;
d, 99 per cent, (i per cent, without) ; e, 99 per cent, (optimum) ; /, 99 per cent,
(some imperfect) ; g, 99 per cent, (imperfect) ; h, 99 per cent, (imperfect).

It is doubtful if any membranes were destroyed by shaking 15 minutes
after their formation (a, b, c). In d 75 per cent, of the membranes were
destroyed.

i These represent a sample of the same lot of eggs tested two hours later
than all other items of this experiment.

FRANK R. LILLIE.

EXP. 3. TEMPERATURE ESTIMATED 15-16° C.

Exposure to Butyric.

Fertilized Unshaken.

Shaken After Membrane
Formation.

Fertilized after Shaking.

a. 90 sees

o segmented

ITT min

i 3 %) segmented

b. 105 "

0

if "

o i<;% "

C. 120 "

0

24- "

o.2S%

a. Membranes generally excentric, many partial.

b. All have fine membranes (optimum).

c. All have membranes, generally good, some excentric.
In a all membranes destroyed ; eggs agglutinated.

In b 70 per cent, membranes destroyed ; much agglutination.
In c 50 per cent, membranes destroyed; much agglutination.

EXP. 7.1 TEMPERATURE 15° C.

Exposure

to Butyric.

Membranes.

Fertilized Unshaken.

a

"?O

sees.

O

95%) segmented

b

4.C

70%

0 2%

c

60

1 4

00%

0

d

7C

II

00%

i effff "

e.

oo

II

on%

/••

GO

4 1

100%

o

g. .

IO?

t 1

100%

o

h....

I2O

4 f

99% (excentric)

o

i

I3S

..

08%

o

j. . .

ISO

t 4

1 00% ? (excentric)

o

k

180

I t

100% ?

o

1

240

II

100% f

o

m

^6O

44

00% "

o

n.
o..

60O
1200

1 4
It

It 4 t

10%

o
o

i This experiment was run in two sections, a-e and f-o. The exposure
times y, e and / were thus repeated.

EXP. 8. TEMPERATURE 17° C.

Exposure to Butyric.

Fertilized Unshaken.

Shaken After
Membrane Formation.

Fertilized after
Shaking.

a. . .
b, . .
c.. . .

75 sees.
90 "
105 sees.

No cleavage

41 II

2.5 min.
3-5 "
4 "

0-3% segmented

0.15%
o

All of a, b and c formed membranes.

Shaking destroyed 50 per cent, membranes in each case.

STUDIES OF FERTILIZATION.

EXP. 9. TEMPERATURE 16.5° C.

Exposure to Butyric.

Shaken After Membrane
Formation.

Fertilized after Shaking.

a.

120 sees.

2 min.

4% segmented

b.

135 "

2 min. 35 sees.

i-5%

c

ISO "

X " 20 "

0.3%

In a only 50 per cent, formed full membranes (under-exposure).

In b 95 per cent, formed full membranes.

In c 50 per cent, formed full membranes (over-exposure).

In a shaking destroyed all memberane.

In b shaking destroyed 50 per cent, of the membranes.

In c shaking destroyed 20 per cent, of the membranes.

LITERATURE.

Loeb, J.

'13 Artificial Parthenogenesis and Fertilization (p. 234). University of
Chicago Press, Chicago, 312 pp.

'15 On the Nature of the Conditions which Determine or Prevent the En-
trance of the Spermatozoon into the Egg. American Naturalist, Vol.
49, PP. 257-285.
Just, E. E.

'19 The Fertilization Reaction in Echinarachnius parma. III. The Nature
of the Activation of the Egg by Butyric Acid. BIOL. BULL., Vol. 36,

PP. 39-53-
Moore, C. R.

'16 On the Superposition of Fertilization on Parthenogenesis. BIOL. BULL.,
Vol. 31-

THE METABOLIC GRADIENTS OF VERTEBRATE
EMBRYOS. I. TELEOST EMBRYOS.

LIBBIE H. HYMAN,
HULL ZOOLOGICAL LABORATORY, UNIVERSITY OF CHICAGO.

I. INTRODUCTION.

The present paper is a record of observations on the disintegra-
tion gradients of some teleost embryos. Since these gradients
have served to explain much that was formerly obscure in the
physiology and development of organisms, it has seemed to us
worth while to study them in as wide a range of living forms,
both adult and embryonic, as practicable. In particular it has
seemed desirable to determine the susceptibility gradients in the
vertebrates in order to analyze their role in normal and terato-
logical development and in the physiology of at least certain
systems.1 Since it is impracticable to study adult vertebrates by
the susceptibility method owing to the difficulty of determining
the time of death of different regions, our attention has been
necessarily confined to embryonic stages. The recent publication
of Bellamy ('19) on the frog represents the first of these in-
vestigations on the vertebrate embryo. In this paper Bellamy
described the disintegration gradients of the frog at various stages
of development and showed that the teratological forms experi-
mentally produced by him and by a number of previous investiga-
tors can be easily and simply accounted for on the basis of these
gradients. Similarly in these papers I shall describe the disin-
tegration gradients of other vertebrate embryos and attempt to
correlate them with well-known facts of normal and teratological
development.

When an organism is exposed to a lethal concentration of a
toxic solution, it is found that all of its parts do not die simul-

1 In particular the gradients have been of great service in interpreting the
physiology of the nervous system, the digestive system, and the heart.

32

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 33

taneously, but that they die at different time intervals. In other
words, the parts of the organism exhibit a differential suscepti-
bility towards the toxic substance. This differential susceptibility
bears a direct relation to the organization and axiation of the
organism. In the simplest cases this relation is : that the anterior
or apical end is the most susceptible and dies and disintegrates
first, and that the susceptibility decreases in a graded manner
from the apical end along the antero-posterior or apico-basal axis.
A death or disintegration gradient (so-called because the time of
death is recognized by the disintegration of the part affected) is
thus observable. This simple gradient has been designated by us
the primary gradient. In more complex forms secondary gradi-
ents often appear after the early stages of development and may
persist throughout life; they consist in the appearance of regions
of high susceptibility other than the apical or anterior end. In
complex animals it is further found that organs and systems may
develop gradients of their own which do not necessarily corre-
spond to or bear any relation to the more general primary gradient
of the organism. As an example of an organ possessing an
axiation independent of that of the rest of the body may be
mentioned the vertebrate heart.

The primary gradient is initiated in protoplasm through the
action of the external world upon it. Such environmental action
is designated a stimulus and the point of stimulation becomes ipso
facto the region of high susceptibility from which both the effect
of the stimulus and the susceptibility gradually diminish, finally
dying out. Under repeated stimulation such gradients, originally
temporary, become morphologically fixed in the protoplasm to a •
more or less permanent degree. Since these gradients arose in
response to the external world they are retained most completely
and in a least altered condition in the external surface of organ-
isms, and its derivative, the nervous system. As the internal
organs and systems increase in number and complexity it is not
to be expected that they would necessarily develop gradients
corresponding to the gradient of the surface structures but rather
that, since the gradient is the expression of physiological activity,
their gradients will be the expression of their relation to the

34 LIBBIE H. HYMAN.

organism as a whole and to its parts. Consequently secondary
gradients are of common occurrence in complex animals.

Owing to the fact that the disintegration gradients are observ-
able and identical in a wide range o*f substances of very varied
chemical properties and constitutions, it is certain that they are
not due to any specific action of these substances upon proto-
plasm. This is further evidenced by the fact that organisms dis-
play the same differential susceptibilities to extremes of physical
conditions, such as high temperatures, and to low oxygen supply.
Owing to the fact that differential susceptibility is directly re-
lated to physiological conditions such as age, starvation, regenera-
tion, motor activity, stimulation, etc., it is further certain that
differential susceptibility is not primarily an expression of struc-
tural or morphological gradations along the axes of organisms.
All of the facts at hand lead us irresistibly to the conclusion that
the gradients are physiological in nature, that they are mani-
festations of a quantitative gradation in function and metabolism
along the axes of organisms. We are therefore accustomed to
refer to these gradients as metabolic gradients ; physiological
gradients would possibly be a better term, as Professor Child has
recently suggested.1

The susceptibiltiy method is thus a method for determining in
a general way the relative rates of activity of different parts of
the organism. Its results do not necessarily correspond to de-
terminations of total metabolism since usually the susceptibility
of only certain parts of the organism can be determined. It is
not claimed that it is an accurate measure of metabolic rate, since
too many factors enter into differential susceptibility, or that it
should supplant methods of directly measuring metabolic rate,
such as oxygen consumption, carbon-dioxide production, or de-
termination of other metabolic end products. It has, however,
served to reveal facts not discoverable by any other method at
present known to us.

i These matters are discussed at greater length in a paper by Child now
in press in the BIOLOGICAL BULLETIN and in a book " The Origin and Devel-
opment of the Nervous System," shortly to be published by the University of
Chicago Press. I have had the privilege of reading the manuscript of both
publications and wish to acknowledge my indebtedness to them for the present
argument.

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 35

In studying the metabolic gradients by the direct susceptibility
method, the observer watches the time of death of different levels
of the organism. It is evident that the completeness and accuracy
of his observations depend on two factors : first, the ease with
which the death point can be recognized, and second, the visi-
bility of various parts of the organisms. In regard to the first
matter, it may be said that the death point is usually recognized
by certain changes in the appearance of the part under observa-
tion ; it becomes white and opaque and loose and finally expands
into a shapeless mass of granules. In the case of vertebrate
embryos, owing to their delicate, almost diaphanous, structure,
these changes are not always detectable with certainty, especially
in very early stages, and repeated observation has often been
necessary in order to make certain of the course of disintegration.
As concerns the second matter, it should be perfectly obvious,
that the gradients can be determined only for those parts of the
organism which can be seen clearly ; in general these will be the
superficial parts. It is usually impossible to observe the death of
the entodermal structures ; such is the case in the present study.
The transparency of embryos, however, permits more extended
observations on internal structures than is usually possible with
adult organisms.

II. DISINTEGRATION GRADIENTS OF FISH EMBRYOS.

i. Material and Method. — The disintegration gradients were
investigated in the embryos of three species of fish — the killifish,
Fundulus heteroclitus, the cunner, Tautogolabrus (Ctenolabrus*)
adspersus, and the cod, Gadus morrhua. They were obtained at
Woods Hole, Mass., the first two species in June and July, and
the cod in December, 1919. Sexually mature Fundulus and
Tautogolabrus were obtained through the supply department and
were stripped " dry," that is into vessels containing no or very
little water. After fertilization had occurred water was added.
In this way a high percentage of developing eggs is obtained, as
previous investigators have found. The eggs of the cod were
obtained from the Bureau of Fisheries at Woods Hole and I am

36 LIBBIE H. HYMAN.

greatly indebted to the director of the fisheries station for his
kindness in supplying me with an abundance of material. Cod
eggs in all stages of development were obtainable at any time.

In studying the gradients the eggs were placed in watch glasses
which were filled with the toxic solution, and observed under the
low power of the compound microscope. Sometimes the watch
glasses were filled full and covered with a circular piece of thin
glass with the exclusion of air. It was found, however, advisable
in most cases to leave the watch glasses uncovered since it was
frequently necessary to turn the eggs with a needle in order to
bring all parts into view. The eggs of the cunner and the cod
are pelagic floating eggs but very soon after being placed in the
solution they become opaque and sink to the bottom. Since the
eggs in all three species rest upon the bottom of the watch glass
it is easy to turn them into any desired position by mean of a.
needle.

For the disintegration of the embryos of Tautogolabrus, solu-
tions of potassium cyanide in sea-water were employed, in con-
centrations of i/ioo mol. or stronger. When work was begun on
the eggs of Fundulus it was immediately discovered that potas-
sium cyanide was useless for the purpose. Apparently the egg
membranes of this fish are impermeable to cyanide for the
embryos will live and their hearts will continue to beat for very
long periods in relatively concentrated cyanide solutions. It was
found that anaesthetics and acids would penetrate the eggs of
Fundulus very readily and stop the heart beat within a few
minutes ; but in none of these substances could the disintegration
gradients be observed for they seemed to coagulate the embryo
and it was consequently impossible to determine the time of
death of different levels. Finally after fruitless trials of many
substances, it occurred to me to try ammonium hydroxide, as it
is well-known that this substance penetrates invertebrate eggs
readily. It proved to be entirely satisfactory and was thereafter
exclusively employed for the study of the gradients of Funduhis
and the cod. In making ammonium hydroxide solutions several
drops of the pure concentrated solution were added to about fifty
c.c. of sea-water ; the sea-water was then shaken thoroughly and

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 37

filtered to remove the abundant precipitate which is always
formed on the addition of alkali to sea-water.

The drawings illustrating the disintegration of fish embryos
were copied from free-hand sketches made while the embryos
were under observation. They do not pretend to be accurate
drawings of the embryos but are simple diagrams. The process
of disintegration is represented by stippling.

2. Disintegration of Early Blastoderm Stages. — The very early
cleavage stages were not investigated. The study began at the
stage -when a small blastoderm is present consisting of a number
of cells. At this time in both Tautogolabrus and Fundulus, the
central cells are found to be the most susceptible and from them
disintegration extends to the periphery of the blastoderm. Figs,
i and 2 show the course of disintegration in the egg of Tauto-
golabrus in an early blastoderm stage consisting of about thirty
cells. The most central cells become crenated at their margins
and eventually resolve into droplets ; this process spreads to the
periphery. Similar observations were made upon Fundulus eggs
but they were not entirely satisfactory. In the ammonium hy-
droxide solutions the blastoderm of Funditlns invariably shrinks
and arches up on the yolk. After this change has occurred, the
central region of the blastoderm bursts open and subsequently
disintegrates ; later this disintegration extends to the periphery.
These changes are illustrated in Figs. 3 and 4. The alkaline
solution disturbs the normal tension existing within the blasto-
derm and between the blastoderm and the yolk ; the blastoderm is
loosened at its periphery and shrinks into a mass much smaller
than normal. This occurrence has rendered it difficult or im-
possible to determine the course of disintegration in the early
stages of Fundulus. It is a question whether such changes as
can be observed, such as those represented in Figs. 3 and 4, are
real expressions of differential susceptibility or whether they may
not be due to the alterations of tension just described. In view,
however, of the observations on Tautogolabrus, about which there
cannot be any question, it is reasonable to believe that in Fundu-
lus also the central region of the blastoderm is more susceptible
than the periphery.

38 LIBBIE H. HYMAN.

In the case of the cod, however, the state of affairs is the
reverse. The earliest stages examined were those consisting of
a small blastoderm composed of a considerable number of cells.
At this time the periphery is more susceptible than the central
portions. This is the case before the germ ring has become
visible. The peripheral portion of the blastoderm becomes
sharply separated from a central area and disintegrates long
before the latter. This condition is illustrated in Fig. 5, the
peripheral ring having disintegrated leaving a central disc still
intact. In very young blastoderms, the periphery is equally sus-
ceptible at all points, as in Fig. 5, but very soon a differential
susceptibility appears. One region of the periphery -is found in
the majority of cases to be more susceptible than any other part
of the blastoderm and from this point of high susceptibility the
disintegration proceeds in both directions along the periphery.
Two stages in the disintegration of such a blastoderm are repre-
sented in Figs. 6 and 7. There cannot be any reasonable doubt,
in view of the conditions at later stages, that this peripheral area
of high susceptibility is the place at which the embryonic shield
is to arise. It may therefore be stated that this region is physio-
logically different from the rest of the blastoderm long before its
morphological role becomes apparent.

The metabolic conditions during the early blastoderm stages of
these three species of fish are therefore of two kinds. In the
case of Tautogolabnis and Fundulus, the central regions are
more susceptible while in the cod the peripheral region has the
highest susceptibility. There can be little doubt that these differ-
ences are the expressions of real differences in the physiology of
development of the two classes of eggs. This matter is discussed
further below.

3. Disintegration of Later Stages of the Blastoderm. — As the
blastoderm of Tautogolabnis expands over the yolk, the region of
high susceptibility comes to lie more posteriorly, that is, in that
portion in which the embryo is to appear. The posterior half of
the blastoderm is markedly more susceptible than the anterior
half, as shown in Fig. 8. This whole region appears as far as
could be determined to be equally susceptible throughout; from

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 39

this region disintegration extends forward along the margins of
the blastoderm. In later stages when the germ ring is approach-
ing the equator of the egg, the region of high susceptibility is
shifted still more posteriorly. At this time the observations are
somewhat obscured by the occurrence already described for
Fundulus. The blastoderm, which covers nearly half of the
egg, as shown in Fig. 9, breaks loose at its periphery and shrinks
and arches up from the yolk, as in Fig. 10. Nevertheless the
course of disintegration was followed in a number of individuals.
A certain area along the margin of the blastoderm is more sus-
ceptible than any other region, as illustrated in Fig. 10; from
this area disintegration spreads forwards and laterally as in Fig.
ii. There cannot be any doubt that the region of highest suscept-
ibility is the place where the embryonic axis is to arise. In
Tautogolabrus by the time that the blastoderm has spread nearly
half way around the yolk, the eggs float in a tilted position so that
half of the blastoderm is on the upper and half on the lower
side. The embryo arises in the center of the lower half. It is
therefore a simple matter to determine in eggs of the stage de-
picted in Fig. 9 where the embryo is to arise. The place where
this occurs is the region of the high susceptibility shown in
Fig. ii.

Observations on the late blastoderm stages of Fundulus were
impossible owing to the behavior of the blastoderm. It bursts at
some point opposite the place where the embryonic shield is form-
ing; it then shrinks rapidly forming a mass about the embryonic
shield. In spite of repeated attempts it was impossible to come to
any conclusions regarding the susceptibility relations at these
stages owing to this shrinkage.

In the eggs of the cod, the germ ring is differentiated at a very
early stage. The germ ring is always more susceptible than the
central part of the blastoderm and one region is more susceptible
than the remainder of its circumference. The disintegration is
therefore the same as before the germ ring has become visible
and as depicted in Fig. 6 and 7. There can be no reasonable
doubt that the point of high susceptibility in the germ ring is the
place where the embryonic shield originates. The shield soon

4O LIBBIE H. HYMAN.

makes its appearance as a slight bulge on the germ ring as shown
in Fig. 12. At this time the region of the shield is the most
susceptible part of the embryo and from this region disintegration
spreads in both directions along the germ ring, as in Figs. 13 and
14. As the embryonic shield grows forward its anterior margin
is most susceptible and from this area disintegration extends
posteriorly in the shield as shown in Figs. 15 and 16.

4. Disintegration of the Early Embryonic Axis. — After the
germ ring has advanced half way or more over the yolk, the
embryo appears in the center of the embryonic shield. Two faint
lines outline its axis. In Tautogolabrus and Fnndulus the em-
bryonic shield is very faintly defined but in the cod is sharply
marked out on the blastoderm. In the two first-named species,
observations on the disintegration of the early embryonic axis
were much obscured by the occurrence of the shrinkage already
described. In but one or two cases in each species was the dis-
integration followed with certainty. In Tautogolabrus at the
time when the embryonic axis first becomes visible, the course of
disintegration is the following; the embryonic axis fades from
view and the region in which it lies undergoes disintegration.
Fig. 17 represents the normal blastoderm at this time together
with the yolk ; the embryonic side of the blatoderm is on the under
side of the yolk. Fig. 18 shows the same blastoderm much
shrunken, the yolk being omitted ; the area where the embryonic
axis was is seen to be in process of disintegration. As far as
could be determined the whole embryonic region is equally sus-
ceptible. In Funduhis an early stage of the embryo was observed
in but one case. In this individual two regions of high suscepti-
bility were observed, one at the posterior end, the other at the
anterior end of the embryonic axis. The first region was the
more susceptible of the two ; from these two points, disintegration
proceeded towards the middle of the axis. Three stages in the
disintegration of this individual are illustrated in Fig. 19.

Early stages of the cod embryo were observed with ease. Dis-
integration begins at the anterior end of the embryonic shield and
proceeds posteriorly, more rapidly at the margins, as shown in
Figs. 20 and 21. In a slightly later stage, the susceptibility

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 41

gradient is the same except that a secondary region of high sus-
ceptibility is faintly evident at the posterior margin of the shield.
This stage is depicted in Figs. 22 to 26.

5. Disintegration of Later Stages of the Embryo. — After the
embryo had become established its disintegration gradient was
observed with ease in all three species and in a great many indi-
viduals. In Tautogolabrus disintegration begins at the anterior
end of the embryo and proceeds posteriorly along the neural tube
to its posterior end. An early embryo is depicted in Figs. 27
and 28 and a later one in Figs. 29 to 32. The disintegration
gradient remains the same up to the time when the germ ring
closes. The eyes are not very highly susceptible but disintegrate
at about the time when the disintegration in the neural tube has
extended half way back. The neural tube usually separates from
the rest of the embryo and becomes arched. The time of death
of the somites could not be determined with certainty as they do
not seem to undergo disintegration but remain distinct long after
the disintegration of the neural tube is completed. After the
germ ring has closed a secondary region of high susceptibility
appears at the posterior end of the embryo as shown in Figs. 33
and 34. From this time on there is no further change in the
disintegration gradients ; there are always two regions of high
susceptibility, one at each end of the embryo ; disintegration pro-
ceeds posteriorly along the neural tube and anteriorly to a slight
extent from the end of the tail. In no case was the fate of
the somites determined.

In Fundulus embryos the disintegration gradients are in general
the same as in Tautogolabrus with certain exceptions. In Fundu-
lus the two regions of high susceptibility are present from th»
earliest observable stages of the embryonic axis. The posterior
end of the embryo is the more susceptible and disintegration begins
there, and progresses anteriorly. Disintegration then begins at
the anterior end of the embryo and progresses posteriorly. In
the very earliest stages as in Fig. 19 this anterior disintegration
begins at the tip of the neural axis. Very soon, however, the
optic bulbs make their appearance. As soon as this has occurred,
the optic bulbs are decidedly the most highly susceptible parts of

42 LIBBIE H. HYMAN.

the anterior end. After they have disintegrated, disintegration
begins at the tip of the forebrain and proceeds down the axis,
meeting the disintegration progressing forwards from the pos-
terior end in about the region of the hindbrain. An embryo of
this stage, at the first appearance of the optic bulbs, is represented
in Fig. 35. The first drawing in this figure shows the normal
embryo, the germ ring not yet closed and somites not yet formed ;
the other three drawings give stages in the disintegration. Both
neural tube and mesoderm are involved in the disintegration
although naturally the fate of the mesoderm is less readily
ascertained, owing to its loose structure.

In later stages of Fundulus embryos the disintegration gradi-
ents remain the same except that the secondary region of high
susceptibility at the posterior end gradually decreases in import-
ance. By the time that a number of somites have formed the
eyes are the most susceptible region of the embryo. Following
the eyes, the forebrain disintegrates and disintegration proceeds
posteriorly. The posterior end then begins to disintegrate and
disintegration proceeds slightly forwards from this region, meet-
ing the other wave of disintegration near the posterior end of the
embryo. The disintegration of an embryo of this age is shown
in Fig. 36. The first drawing shows the normal embryo, the other
four the course of the disintegration. The shrinkage of the
embryo accompanied by a sinuous bending of the neural tube,
which always occurs in later embryos in the killing solution, is also
illustrated. The fate of the somites could not be observed with
certainty and hence they are omitted but the high susceptibility of
the segmental plate region is shown in the second drawing of
Fig. 36. In still later stages of Fundulus embryos, the eyes are
no longer more susceptible than the forebrain but both disinte-
grate about simultaneously. The fate of the somites could not
be observed with certainty as they do not disintegrate readily. In
earlier stages the somites appear to disintegrate from the posterior
end forwards ; in later stages when a number of somites have
appeared the anterior and posterior somites seemed to be the most
susceptible, the middle ones less susceptible.

The latest stages of Fundulus which were investigated are de-

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 43

picted in Fig. 37. The disintegration gradient is much the same
as described in the preceding paragraph except that a region of
high susceptibility has developed in the hindbrain where the
cerebellum is forming. The end of the tail has by this time
become free from the yolk and is elongating; its susceptibility is
increased relative to the preceding stage. The tip of the fore-
brain and the eyes are about equally susceptible and from them
disintegration proceeds posteriorly along the axis in both neural
tube and mesoderm. The most anterior and posterior somites
are more susceptible than the others and from them disintegra-
tion proceeds in both directions to the middle of the embryo.
Investigation of later stages of Fundulus embryos was not feas-
ible as the embryos do not disintegrate as readily as previously.
It may be stated, however, that the auditory vesicles and the buds
of the pectoral fins were observed to be regions of high sus-
ceptibility.

The later stages of the cod embryo resemble those of Fundulus
except with reference to the eyes. There are always two regions
of high susceptibility, the anterior and posterior ends of the
embryonic axis. The anterior end commonly precedes in stages
before the closure of the germ ring. Disintegration begins at the
tip of the forebrain and passes backwards along the brain and
eyes; next the germ ring at the posterior end of the embryonic
axis disintegrates and this disintegration extends forwards ; the
two waves of disintegration meet anterior to the middle of the
embryo. The eyes are not highly susceptible as in Fundulus but
about as susceptible as the forebrain. This stage is illustrated in
Fig. 38. As the germ ring closes the susceptibility of the pos-
terior end increases and the eyes become temporarily more sus-
ceptible than the forebrain, as in the stage depicted in Fig. 39,
where the germ ring is on the point of closure. A later stage is
illustrated in Fig. 40. The susceptibility of the eyes has de-
creased but otherwise the disintegration gradient is the same as
in the preceding stage. The disintegration of the somites as
shown in Figs. 39 and 40 is the same as in Fundulus, the disin-
tegration proceeding from each end to the middle. In both
- 39 and 40 a region of high susceptibility exists at the level

44 LIBBIE H. HYMAN.

of the anterior end of the hindbrain ; this appears to be related to
the development of the auditory vesicles. In later stages evi-
dences of the appearance of a region of high susceptibility asso-
ciated with the formation of the cerebellum are present.

6. The Gradient of the Fundulus Heart. — This investigation
was in reality undertaken for the purpose of studying the gradient
of the heart. I was convinced from the facts known about the
heart that such a gradient must exist. My observations show that
such is the case ; however, the matter proved more difficult of
demonstration than was anticipated owing to the fact that the
outer surface of the heart does not readily undergo disintegra-
tion. It was only after repeated observation that it was dis-
covered that the disintegration processes occur only in the interior
of the heart, the surface outlines remaining intact.

The gradient was studied in the Fundulus heart only, the hearts
of the other two species of fish being too small for the purpose.
The early stages of Fundulus proved unfavorable as the yolk
sac bulges up around the embryo and conceals the heart from
view. In later stages, after the heart beat is established, the dis-
integration gradient is invariably as follows. Disintegration
begins in the wall of the sinus venosus and progresses rapidly
along the heart tube to the arterial end of the heart. In this dis-
integration, the heart wall dissolves or melts away, leaving how-
ever, the external outlines intact ; this process sweeps rapidly
along the heart from the sinus to the bulbus arteriosus. This
gradient was observed in numerous cases in hearts in which there
was no visible differentiation along the cardiac tube. In later
stages after such differentiation has occurred the gradient is more
marked and steeper, that is, the disintegration passes more slowly
along the heart. After the interior has disintegrated the heart
tube expands but the outlines remain intact.

The gradient in the heart is shown not only by the course of
the disintegration but also by the order in which the chambers
of the heart cease beating in toxic solutions. In younger hearts,
the order of cessation of beat is : bulbus and ventricle, auricle,
sinus venosus. The sinus continues to beat feebly after the
other parts of the heart have stopped contracting. The explana-

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 45

tion of this is that the heart beat originates in the sinus venosus,
as has long been known, and is transmitted from the sinus in
sequence along the heart tube ; in order that such transmission
shall occur, the original impulse must attain a certain strength.
As the sinus is the most susceptible part of the heart, its beat is
weakened by the action of the toxic solution and the impulse
generated in it becomes too feeble to be transmitted along the
length of the heart tube. At first it is able to reach as far as the
auricle but not to the ventricle and bulbus, which consequently
no longer contract while the auricle still continues to beat; sub-
sequently the beat becomes too feeble to be transmitted as far as
the auricle and the sinus remains beating by itself. In later
stages of the heart, the auricle may continue to beat after the
sinus has ceased ; it is probable that as development proceeds, the
auricle develops some slight degree of independence and auto-
maticity of its own, as is well known for the auricles of the
lower vertebrates, and being less susceptible to toxic agents than
the sinus it may continue to contract after the latter has ceased.

The heart is by far the most susceptible part of the Fundulus
embryo and dies and disintegrates shortly after the embryo is
exposed to the ammonia solution. Older hearts are more sus-
ceptible than younger ones. This indicates that the metabolic
rate of the heart increases during development.

These observations show that there is a gradation in meta-
bolic rate along the heart tube from the sinus to the arterial end
of the heart. Such a gradation is in all probability the cause of
the sequence of the heart beat. The sequence of the heart beat
is generally stated in textbooks of physiology to be due to the
fact that the venous end of the heart possesses a more rapid
intrinsic rhythm than the other chambers of the heart and con-
sequently " sets the pace " for them. This is really only another
way of saying that the venous end of the heart has a higher rate
of metabolism than the rest of the heart tube, for how could it
contract more rapidly than they if such were not the case?
Nevertheless it does not seem to have occurred to physiologists
that in such a simple gradation in metabolic rate rests the ex-
planation of the sequence of the beat. It is a familiar physio-

46 LIBBIE H. HYMAN.

logical fact that the more rapidly an organ is respiring, the more
rapidly it gives off carbon-dioxide and consumes oxygen, the
more rapidly does it function. A case very similar to the heart
is that of the digestive tract, in which Alvarez ('18) has shown
that the duodenal end of the intestine has the highest irritability,
fastest respiratory rate, most rapid rate of contraction, and
greatest susceptibility to drugs, of any part of the intestine and
that these factors decrease along the intestine. The matter of
the cause and sequence of the heart beat will be discussed more
fully in later papers, as the gradient is more easily demonstrable
in the chick heart.

7. General Summary of Observations on the Gradients of
Teleost Embryos. — Before the appearance of the embryonic axis
the central region of the blastoderm is more susceptible in Fundu-
lus and Tautogolabms, the peripheral region in the cod. It is
highly probable that the central region would also be found to be
more susceptible in the cod if a sufficiently early stage were in-
vestigated; but unfortunately this was not done. It is evident,
however, that the high susceptibility of the central region, if ever
present in the cod, is lost at a very much earlier stage than in the
other two species. In Tautogolabms (Fundulus being unfavor-
able for observations) the region of high susceptibility then
gradually shifts from the central to centre-posterior regions and
finally to a point on the germ ring where the embryo is to appear.
In the cod also the region of high susceptibility becomes limited
to the region of the germ ring where the embryonic shield subse-
quently develops. From this place in both species the region of
high susceptibility grows forwards simultaneously with the ap-
pearance of the embryonic axis. In the embryos of all three
species of fish there are sooner or later two regions of high sus-
ceptibility, the anterior and the posterior end, from both of which
disintegration extends towards the middle. The posterior region
of high susceptibility arises very early in Fundulus, later in the
cod, and very late in Tautogolabms. This double gradient exists
in both nervous and mesodermal structures, but nervous struc-
tures are as a rule far more susceptible. A very high suscepti-
bility of the eyes is a marked feature of Fundulus embryos.

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS, 47

Regions of high susceptibility also appear in connection with the
development of the auditory vesicles and the cerebellum. A
gradient exists in the heart (Fundulus}, the venous end being the
most susceptible and the susceptibility gradually decreasing to the
arterial end.

III. RELATION OF THE GRADIENTS TO NORMAL TELEOST

DEVELOPMENT.

While it is not my purpose to enter into a detailed account
of teleost development it may not be amiss to point out the bear-
ing of the observations recorded in the preceding section on the
normal course of development. In discussing this matter it is
necessary to bear in mind the fact that teleost development is
highly specialized. Embryologists agree in general that the
primitive mode of vertebrate development is illustrated by the
ganoids and amphibia, forms with total unequal cleavage. The
teleost mode of development is derived from some such type and
has probably proceeded along different lines in different groups
of teleosts.

In the early stages of the blastoderm it was found that in
Fundulus and Tautogolabrus the central cells have a higher rate
of activity, as measured by the susceptibility method, than the
marginal cells, while in the cod the margin of the blastoderm is
more active. These facts indicate that we are dealing here with
two different modes of development. In the two first-named
species the germ ring and the embryonic shield are poorly differ-
entiated and only faintly visible ; in fact I have not represented
them in my figures on this account. It is highly probable that
in these species the germ ring plays a minor role in the formation
of the embryo and that the embryo is produced chiefly through
the activity of the central region of the blastoderm. This con-
clusion is supported by the observations and experiments of
Morgan ('95) on Tautogolabrus and of Sumner ('03) on Fundu-
lus. Morgan concluded that the embryo of Tautogolabrus is
formed largely of material that has never been part of the germ
ring. He also showed that the development is not affected if cuts
are made in the germ ring at each side of the early embryonic

48 LIBBIE H. HYMAN.

shield. Sumner concluded that in Fundulus the expansion of the
blastoderm is centrifugal since when needles are inserted in the
sides of the blastoderm the germ ring fails to expand at the
points adjacent to the needles and bays in the germ ring conse-
quently appear at those places. Similar conditions seem to be
present in the salmon since according to the experiments of
Kopsch ('96) the destruction of spots in the germ ring at each
side of the earliest stage of the shield does not affect the forma-
tion of the embryo. It is evident that this type of development in
which the center of the blastoderm is the region of high activity
harks back to such a condition as that seen in the ganoids and
amphibia in which the animal pole is the region of greatest ac-
tivity, as evidenced by more rapid rate of division and greater
susceptibility to toxic agents (Bellamy on the frog).

In the cod, on the other hand, development proceeds in a dif-
ferent manner. The expansion of the blastoderm over the yolk
is here evidently largely due to the activity of the germ ring. In
this fish as contrasted with the other two species, the germ ring
becomes visible at a very early stage and it and the embryonic
shield are very sharply marked off from the rest of the blasto-
derm. My observations further show that the margin of the
blastoderm in the cod is already differentiated as a region of
high activity before the germ ring is morphologically distinguish-
able. The mode of development of the cod by peripheral ex-
pansion may be regarded as a specialization from the more
primitive type exhibited by Tautogolabrus and Fundulus and
probably represents a short cut in development with omission of
the original centrifugal method of growth.

With the expansion of the blastoderm we next observe in Tau-
togolabrus, a shifting of the region of high activity from the
central region of the blastoderm to the central-posterior and
finally to a definite region of the germ ring where the embryo is
to form. It seems to me probable that in these changes in the
Tautogolabrus blastoderm we have an illustration of the manner
in which the centrifugal method of development is transformed
into the germ ring type. The conditions in Tautogolabrus thu*
eventually come to be identical with those very early present in

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 49

the cod. The cod has evidently omitted the early stages and very
soon arrives at the condition in which the posterior median region
of the germ ring is the center of activity. There has thus been
produced the typical teleost method of development by means of
the embryonic shield.

In vertebrates like the frog having a primitive mode of de-
velopment it is important to note that a region of high activity
also develops in the median posterior point of the ' germ ring,"
that is to say, the dorsal lip of the blastopore (Bellamy, loc. cit.}.
The frog, however, also retains the region of high susceptibility
at the animal pole, so that two regions of high activity are con-
stantly present — the animal pole and the dorsal lip of the blasto-
pore. In the teleosts the former region either is not present from
the first or first develops and is then lost ; but subsequently a new
region of high activity develops at this point. It seems therefore
that the germ ring type of development is highly specialized and
the attempt to interpret the modes of development of other verte-
brates in terms of the germ ring is a mistaken effort ; but rather
germ ring types should be interpreted as modifications of the
method illustrated by the amphibia.

After the region of high activity has become limited to a
median posterior area of the germ ring, it extends forwards and
the embryo appears in its center. That the germ ring is actively
concerned in this extension appears improbable since the an-
terior end of the shield in the cod and all of the shield except the
place where it meets the germ ring in the cunner are the regions
of high susceptibility. If the forward growth were produced by
the germ ring, by a sort of pushing process, then one would ex-
pect the posterior end of the shield to exhibit the highest sus-
ceptibility. It seems probable that the shield grows forwards
through the activity of its anterior end or possibly in some species
with the aid of the surrounding cells of the blastoderm. The con-
ditions are evidently variable in different species. According to
Kopsch ('96) if the embryonic shield of the salmon embryo is
destroyed the germ ring continues to close but no embryo is
formed. Here evidently the embryo arises solely from the ma-
terial of the shield. In Tautogolabrus and Fundulus, however, the

50 LIBBIE H. HYMAN.

cells around the shield probably take part. Thus Morgan ('95)
found that in Tautogolabrus the embryo is formed by a concen-
tration of material towards the center of the shield. Sumner
('03) states that if the embryonic shield of Fundulus is destroyed
a new embryonic shield is regenerated. My observations also
show that in Tautogolabrus and the cod the region about the
shield takes part in the formation of the embryo since a large area
of high susceptibility is present in the early stages, as seen in my
Figs. 18 and 20 to 26. While materials around the shield may
contribute to the embryo the work of Kopsch, Morgan, and
Sumner agrees in denying such a role to the germ ring, in the
formation of the early embryonic axis. Cuts or injuries in the
germ ring at the sides of the shield do not affect the formation of
the embryo but this proceeds in practically normal manner.

Although according to these lines of evidence the germ ring
plays no role or only a very minor one in the early development of
the embryo, it sooner or later becomes involved in the formation
of the embryo. This is shown in all three species by the appear-
ance of a region of high susceptibility at the posterior end of the
embryonic axis. This active region arises very early in Fundulus
and this fact indicates that the greater part of the Fundulus
embryo is laid down by this posterior growing region. This was
also the conclusion reached by Sumner since he found that pierc-
ing this growing region inhibits the formation of the posterior
part of the embryo. This region, grows backwards adding to the
embryo in front of it, exactly as in the case of the primitive
streak of the chick, with which it is no doubt homologous. In the
cod the posterior growing region arises somewhat later; and in
Tautogolabrus very late. In the latter species it is evident that
very little of the axis of the embryo is due to the activity of this
posterior region and that the embryo elongates considerably with-
out it, apparently through growth at its anterior end. This inde-
pendence of the embryo of Tautogolabrus of the germ ring was
already noted by Morgan in 1895 since he found that by placing
the eggs in diluted sea-water it was possible to retard or prevent
the formation of the embryo without affecting the growth or
closure of the germ ring. Experiments on other species are

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 5!

unfortunately lacking except in the case of the salmon, in which
according to the experiments of Kopsch, most of the trunk and
tail of the embryo are produced through the activity of the por-
tions of the germ ring immediately adjacent to the early embry-
onic shield. It seems in view of the facts at hand legitimate to
draw the conclusion that in different teleost embryos the amount
of material contributed to the formation of the embryo by the
germ ring is variable. At one extreme are cases like Tautogola-
brus in which very little material is so contributed and at the
other extreme the case of Fundulus, where most of the embryo
is formed from a growing point in the germ ring. The disagree-
ment among investigators concerning the mode of formation of
the teleost embryo is evidently due to the circumstance that they
worked with different species. The mode of development in dif-
ferent species is not identical but rather exhibits various degrees
of modification from the primitive vertebrate type illustrated
by the ganoids and the amphibia to the extremely specialized
type in which the germ ring plays the dominant role. Owing to
this fact generalizations cannot be drawn from the study of a
single species and controversies which have arisen in the past
concerning the mode of origin of the teleost embryo are without
point.

While it may be said that the evidence indicates that the germ
ring does add to the embryo in varying degrees in different
species, the facts recorded in this paper do not seem to me to
support the theory of concrescence. The embryo is not produced
by a concrescence of two areas of the germ ring even in cases like
the cod where the germ ring plays an important role in develop-
ment ; but there is present in the germ ring at the posterior end
of the embryonic axis a region of high activity which in some
species, like Fundulus, is of great importance in the formation
of the posterior part of the embryo. Some of the material of
the germ ring passes into the embryo but not in the manner re-
quired by the concrescence theory; and the anterior end of the
embryo is not formed in this way but by an independent region
of activity.

52 LIBBIE H. HYMAN.

The existence in the development of vertebrates — it is now
known to occur in the frog and chick as well as in teleosts — of two
regions of high activity, one at each end of the axis, is correlated
with the fact that vertebrates are segmented animals. This mode
of development is common, as far as our investigations go, to all
segmented animals. This " double gradient," as we call it, was
first discovered in the annelids (Hyman, '16, and Child, '17).
It appears in an early stage of development in annelids and per-
sists throughout life in all of them, so far as tested. It also
appears, as we have seen, in the vertebrate embryo and persists
as long as the posterior end continues to develop and elongate.
In the annelids where new segments form continually throughout
life, the posterior region of high susceptibility is permanent and
never brought under complete control of the anterior end ; but in
the vertebrates it eventually dies away and segment formation
thereupon ceases. The posterior region of activity is therefore
correlated with the process of segment formation. There can be
little doubt that segments represent incomplete individuals. As
Child ('150) has shown, the anterior end of an organismic axis
is dominant over a certain length of the axis ; beyond this level
physiological isolation occurs, the metabolic activity increases, and
new individuals arise. A similar process is at the basis of seg-
mentation. The posterior growing region of the embryos of
segmented animals has escaped from the control of the anterior
end ; it is physiologically isolated, develops a high metabolic rate
and proceeds to the formation of new but incomplete individuals,
that is, segments. This process of segment formation will con-
tinue indefinitely if the anterior end fails to regain control of the
entire length of the axis as in the annelids but will cease when
this occurs as in the vertebrates and probably arthropods. These
matters have been discussed at greater length by Child ('17)
and Bellamy ('19).

IV. RESPIRATORY RATE DURING THE DEVELOPMENT OF

FUNDULUS.

I have made some measurements of the rate of oxygen con-
sumption and the rate of carbon-dioxide output during the de-

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS.

53

velopment of Fundulus heteroclitus. Similar experiments had
previously been performed by Scott and Kellicott ('16), but ap-
parently no report of these experiments beyond the abstract re-
ferred to has ever been published. In this abstract it is stated
that during the early stages of Fundulus up through the forma-
tion of the embryo the rate of oxygen consumption is less than
o.io c.c. per hour per thousand eggs. This statement is in gen-
eral true; however, it misses the really fundamental point about
the respiratory rate during the early stages. I do know why the
marked change which occurs at the time that the germ ring is near
the equator of the egg should have missed the attention of these
investigators; but at least no mention of this is made in their
abstract. They noted a marked increase in rate of oxygen con-
sumption when the heart begins to beat ; beyond this time there is
no marked rise but a general upward trend with a great increase,
after hatching.

TABLE I.

RATE OF OXYGEN CONSUMPTION OF THE EGGS OF FUNDULUS HETEROCLITUS

DURING DEVELOPMENT.

Results given in cubic centimeters of oxygen consumed per 1000 eggs per
two hours in exps. i, 2 and 3 ; per 3 hrs. in no. 4. Temp. 20° C.

No. of E>

periment

i

2

3

No. of Experiment.

4

Time Since Fer-
tilization.

State of Embryo.

Oxygen Con-
sumed.

Time Since Fer-
tilization.

State of
Embryo.

Oxy-
gen
Con-
sumed.

4—6 hrs

2-4 cells
32 cells
many cells
large disk
far over yolk
embryo
present

good circula-
tion

0.07
0.07
O.OQ
0.09
O.I2
0.14

O.IO
O.I?

0.15

O.22

O.09
O.06
O.O9
O.O9
0.13
0.13

O.IO

0.18

0.15
0.23

0.08
O.O7
O.O7
0.08
0.15
O.I4

O.IO

0.17

0.14
0.24

2—? hrs

to 8 cells
small disk
disk \ over
yolk
embryo with
eyes

circulation

O.O4
O.O6
0.19

O.2I

0.22
0.36
O.27
0.25
0.30

6-8 hrs

9-II hrs. . . .
26-29 hrs. . .

34-37 hrs. . .

2\ days ....
3* days ....
4 + days . . .
Si days ....
later
(average)

9—11 hrs

22—24 nrs- • •

30—32 hrs

2 days

3 days

4 days . .

6 days

Later (average)

My determinations of the rate of oxygen consumption were
made by the same method used by Scott and Kellicott. The
carbon-dioxide output was determined by the phenolsulphontha-
lein method. Forty eggs of approximately the same stage of

54 LIBBIE H. HYMAN.

development were placed in tubes with the phenolsulphonthalein
solution in sea-water. The pH of the sea-water at Woods Hole
was found to be 8.2 by comparison with the set of standardized
tubes put out by Hynson, Westcott, and Dunning. The length
of time required for the forty eggs in a closed tube to turn the
indicator from pH 8.2 to pH 7.6 was recorded at different stages
of development. This furnishes a rough measure of the rela-
tive rate of carbon-dioxide output at successive stages. The
temperature was of course kept constant as nearly as practicable
throughout such experiments.

The data on the oxygen consumption are given in Table I.
Four experiments were run, each consisting of over a thousand
eggs mixed from a number of females. Three of these experi-
ments were run simultaneously, the fourth one at a later time.

The experiments recorded in Table I. show that the oxygen
consumption remains about the same through the early cleavage
although a slight rise probably occurs. By the time, however,
that the blastoderm has spread one third or half way over the
yolk a marked rise in the rate of oxygen consumption occurs. In
experiment 4 this rise was over 200 per cent., less in the other
three experiments. From this time on through the establishment
of the embryo the rate remains about the same and may even fall
again. Thus the formation of the embryo is not a period of
increase in the rate of oxygen consumption but rather the time of
high respiratory activity is that period when the germ ring is
approaching the equator of the egg. This probably corresponds
to the time of gastrulation. After the heart has begun to beat
the oxygen consumption rises again as also found by Scott and
Kellicott. Beyond this time the determinations yielded rather
irregular results. I cannot verify the statement of Scott and
Kellicott that there is a general upward trend during these later
stages ; I found a considerable variability in the amount of
oxygen consumed ; in general it was very little if any higher than
the rate at the time the heart had begun to beat vigorously. In
the table the average of these later determinations is given. It
should be emphasized that the determinations during later stages
are probably unreliable owing to the growth of bacteria on the

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 55

eggs. Although the eggs were frequently washed, particularly in
experiment 4, this source of error was probably present. It
tends of course to make the oxygen consumption appear too great.
In experiment 4 in which the eggs were thoroughly washed twice
daily, the results are probably more reliable than in the other three
experiments ; and in this experiment no increase was observed in
later stages.

These experiments lead us to believe that the rate of oxygen
consumption in the development of Funduhis is highest at the
time when the germ ring is in the neighborhood of the equator,
early on the second day of development. It is probably actually
highest per unit weight of protoplasm since from that time on
the amount of protoplasm increases greatly but the oxygen con-
sumption does not increase in like proportion ; in fact, a consider-
able part of the oxygen consumption after the third day is due
to the activity of the heart. As the embryo is continually increas-
ing in size after this time while the oxygen consumption shows
relatively little increase we may reasonably conclude that the
oxygen consumption of the embryo per unit weight is actually
decreasing. In other words, senescence is already in progress.

The study of the carbon dioxide production yielded similar
results. The carbon-dioxide production per unit time increased
up to the early part of the second day of development after which
it fell, rising again in later periods.

This result, that the metabolic activity of the embryo is at its
highest point at the period when the germ ring is near the
equator, is in harmony with and furnishes an explanation of
previously known facts. Child determined the susceptibility of
the eggs of Funduhis to phenyl urethane ('15^, p. 416). He
found that the embryos are killed more quickly at this stage than
at any other stage. Since susceptibility is, as I have pointed out
in the introduction, a measure of metabolic rate, this result of
itself shows that the metabolic rate is highest at that period. My
experiments confirm this result of Child's and further illustrate
the reliability of the susceptibility method as a measure of rate
of activity. Various investigators, as Stockard and Kellicott,
whose work is considered at greater length later, have noted that

56 LIBBIE H. HYMAN.

the Fundulus embryo is most affected by reagents when the germ
ring is near the equator of the eggs and yields the maximum num-
ber of teratological forms at this period. Newman ('15) found
that in heterogenic fish hybrids development very frequently stops
at this stage.

V. RELATION OF THE GRADIENTS TO TERATOLOGICAL

DEVELOPMENT.

The literature on the structure, occurrence, and experimental
production of teratological vertebrate embryos has now attained
such vast proportions that an adequate review of it would be a
huge task. I shall here consider only the experimental produc-
tion of terata among teleosts. The discussion applies, however,
to vertebrate terata in general, since the mode of development is
much the same throughout the vertebrates.

Morgan ('95) found that if the developing eggs of Tauto-
golabrus are placed in diluted sea-water, the development of the
anterior end of the embryo is commonly inhibited ; the medullary
folds fail to close and the anterior end remains flattened out on
the yolk. In one case the formation of the embryo was com-
pletely inhibited but the germ ring continued to develop and close
in normal fashion.

The experiments of Stockard ('06, '07, '09, '10) on the produc-
tion of teratological forms in Fundulus are familiar to every one.
His first experiments were performed with lithium chloride. In
the stronger solutions, the eggs either cease to develop in an early
blastoderm stage or else very abnormal embryos are produced
with poorly developed anterior and posterior ends, short bodies,
and no eyes. The trunk and auditory vesicles are, however,
present. In weaker solutions of lithium chloride, the embryos
are less abnormal. The expansion of the blastoderm over the
yolk may be retarded resulting in spina bifida. The embryos are
commonly short with no eyes or defective eyes. The developing
eggs were found to be most sensitive to the treatment between 18
and 22 hours after fertilization when the germ ring is near the
equator of the egg. In later experiments Stockard found that
a number of substances would produce the same effects on the

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 57

Fundulus eggs. He used potassium chloride, lithium nitrate and
sulphate, calcium chloride, ammonium chloride, and magnesium
chloride, alone or in combinations. In solutions of all of these
substances embryos were obtained with poorly developed heads
and eyes, or with no eyes, with abnormal hearts and defective cir-
culatory systems, with shortened bodies, and open blastopores. In
magnesium chloride in particular, fifty per cent, of embryos with
various eye defects were obtained. All degrees of approximation
of the eyes were noted, to the cyclopean condition. Many one-
eyed monsters were obtained, in which one eye was small or
defective or wanting. Associated with the approximation of the
eyes was often an abnormality of the anterior part of the head
resulting in displacement and elongation of the mouth which pro-
jected ventrally like a proboscis. The forebrain in these embryos
with abnormal eyes may be nearly normal or reduced ; it is always
reduced when the cyclopean eye is reduced and defective. In
the extreme cases, the olfactory pits were fused also. Later
Stockard found that similar conditions could be produced by
anaesthetics, except that the eye defects were then accompanied
by other defects while with magnesium chloride it is possible to
produce defective eyes in embryos otherwise nearly normal. The
embryos produced in anaesthetics in addition to defective eyes
nearly always have narrow and defective brains, abnormal ear
vesicles, and defective posterior ends in the form of spina bifida.
Other investigators have obtained similar results. McQendon
('i2a and b) obtained cylclopic Fundulus embryos by means of a
number of salts, anaesthetics and alkaloids. He states that in
nature cyclopic trout embryos arise in water containing an in-
sufficient quantity of oxygen and that he has observed cyclopean
smelt embryos which were possibly caused by an excessive car-
bon-dioxide content. Gee ('16) obtained abnormal Fund id us
embryos similar to those of Stockard by alcohol and sodium
hydroxide. These embryos were characterized by defective
heads and eyes, asymmetrical eyes, absence of eyes, shortened
bodies, defective circulation, and spina bifida. Gee found that
the defects are obtained if the egg is exposed to the solutions
before fertilization or shortly after fertilization. Kellicott ('16)

58 LIBBIE H. HYMAN.

obtained numerous defective forms in Fundnlus by exposing the
eggs to low temperature at various periods after fertilization.
The eggs do not develop while in the refrigerator but if removed
even after a number of days to room temperature, some of them
will develop and numerous abnormalities are produced. Although
it is stated by Kellicott that every possible abnormality arises
under these conditions, yet perusal of his data show that the.
abnormalities are in fact limited to certain parts of the embryos.
These are : absent or defective head, absent or shortened tail,
various abnormalities of the brain and eyes, abnormalities of the
heart or circulatory system, abnormalities of wandering cells and
their products.1 It is evident that the terata obtained by Kelli-
cott fall under the same heads as those obtained by Stockard
and others. Kellicott noted a marked susceptibility to low
temperature at the time when the germ ring is approaching the
equator. The low temperature used by Kellicott is more effective
than the chemical solutions employed by others since it greatly
inhibits the development without at the same time destroying the
blastoderm. It is important to note that when such greatly in-
hibited living masses are restored to room temperature sugges-
tions of organs develop such as " brain fragments, lenses, portions
of optic cups, groups of somites, masses of erythrocytes, rhythmic-
ally contractile cells arranged either as flat sheets or tubular
hearts, scattered pigment cells of the usual types, endothelial
cells over the surface of the yolk, fragments of notochordal
tissue." Kellicott did not notice the fact that these fragments
which develop from eggs retarded in early stages concern exactly
the same parts of the embryo in general as fail to develop when
the eggs are inhibited at later periods in their development. Loeb
('15) obtained blind Funduhis embryos by means of potassium
cyanide solutions and low temperatures. One embryo is figured
by Loeb which possesses eyes and tail and nothing else. Werber
('16) again obtained the same teratological types with butyric
acid and acetone-embryos with defective heads, including brain
(forebrain), mouth, and eyes, with approximated olfactory pits,

i No observations were made by me on the wandering cells of the yolk sac.
It is reasonable to believe, however, that such cells are cells of high physio-
logical activity and hence highly susceptible to toxic agents.

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 59

defective auditory vesicles, defective or absent tails. Werber also
noted the same fact which as been mentioned in connection with
Kellicott's experiments, namely, that in some cases, the parts
which are usually inhibited may alone survive, the rest of the
embryo having disappeared. Such isolated parts are the anterior
end of the embryo and the eyes. In some cases the only differ-
entiated parts of the embryo were a fragment of the brain with
an eye attached.

Similar terata can also be produced by hybridization. Such
terata in Fund nl its hybrids were described by Newman ('08, '17)
and Loeb ('15). Newman showed that there is a correlation
between the rate of development of such hybrids and the degree
of abnormality. Those which develop most slowly showed the
most pronounced abnormalities. The terata are of the same
types as those already described, consisting of defective and in-
hibited heads, brains and eyes, defective hearts, shortened bodies,
as well as types in which the head and eyes alone are present.

It is highly significant to note that similar terata can be ob-
tained by treatment of the sperm alone. Oppermann ('13) ob-
tained them from normal eggs of the salmon fertilized by sperm
which had been exposed to radium and mesothorium. The
embryos resulting from such fertilizations show all of the typical
defects — distortions and marked inhibition of the forebrain and
eyes and general anterior end of the body, defects or inhibitions
of the tail, spina bifida, some abnormality of the myotomes.
Embryos were frequently obtained having neither definite heads
nor tails, but only trunks. G. and P. Hertwig ('13) treated the
sperm of Gobius jozo with methylene blue and methyl green and
observed that eggs fertilized by such sperm produce abnormal
embryos with defective anterior and posterior ends.

From this consideration of the literature it is obvious that a
large variety of agents and conditions produce the same defects
in fish embryos. These defects are primarily concerned with the
following parts of the embryo : the forebrain, the head in general,
the sense organs, especially the eyes, the heart and circulatory
system, the tail.

The explanations of these defects have been almost as numer-

6O LIBBIE H. HYMAN.

ous as the investigators concerned. These explanations have in
general proved inadequate and unsatisfactory and fail to account
for the facts. The most obvious explanation, proposed at first
by Stockard, that the defects are the consequence of a specific
action of the chemicals employed upon the embryo, was later
abandoned by him and must be regarded as untenable. The fact
that a large number of substances and conditions call forth the
same defects at once shows that their action must be a very gen-
eral one and not at all specific. The osmotic pressure of the solu-
tions cannot be the effective factor, since solutions of varying
osmotic pressure yield similar results. McClendon's proposal
that the solutions alter osmotic conditions in the egg by changing
the permeability of the surface cannot be accepted in view of the
fact that the same defects are produced by injuring the sperm
only and keeping the eggs in normal sea-water. Stockard's final
conclusion that the defects are due to a general depression of the
eggs by the agents to which it is exposed contains part of the
truth but fails to account for the fact that only certain parts
of the embryo are affected. Werber believes that the defects are
due to a blastolytic destruction or dispersal of the embryo ; but
outside of the fact that such blastolysis cannot be demonstrated
the theory fails like the others to account for the differential ac-
tion of the effective agents on the embryo. Kellicott sought the
explanation in a disturbance of the normal organization of the
egg with abnormal arrangements and distributions of the egg
materials. This theory likewise does not account for the dif-
ferential effect on the embryo.

It is perfectly obvious that the outstanding fact which must be
taken into consideration is that all of the reagents and condi-
tions affect some parts of the embryo more than they do other
parts. These affected parts have already been enumerated. It
is quite impossible to account for this except on the assumption
that certain parts of the embryo are more susceptible to altera-
tions of conditions than other parts. The necessity for this as-
sumption has been recognized clearly by Stockard, McClendon,
and Werber, but it does not seem to have occurred to them that
when this assumption is granted no further explanation is neces-

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 6l

sary. It is of itself the explanation. The defects are due not to
the agents used, except in a general way, but to the metabolic
conditions in the egg and embryo.

The work of Child and his students upon the susceptibility
gradients of organisms has shown that in fact some parts of the
organism are more susceptible to external agents than others.
The differential susceptibility required to explain teratological
development is then no longer an assumption but a demonstrated
fact. In the sea-urchin ('16) and in annelids ('17) Child showed
that the development could be controlled and modified on the
basis of the susceptibility gradients and predictable types of
terata experimentally produced. A similar demonstration of the
relation between the susceptibility gradients and the teratological
development was made by Bellamy on the frog.

The relation between the susceptibility gradients and the pro-
duction of terata is the following: Those parts of the egg or
embryo having the highest susceptibility and metabolic rate are
the most strongly affected by altered conditions of a depressing
nature and the most greatly inhibited by them, providing that the
circumstances do not permit of recovery or acclimation. On the
other hand if the circumstances do permit of such recovery and
acclimation than those same parts which under more severe condi-
tions succumb are able to recover and continue to develop while
parts of lower metabolic rate cannot.

In order to apply these conceptions to any particular organism
it is first necessary to study the metabolic gradients in that organ-
ism. This I have done in the case of the teleost fishes and I have
shown that the most susceptible parts are the forebrain, the eyes
(particularly in Fundulus}, the heart, the posterior end, and to a
less extent the other sense organs.1 It will be obvious without

1 No observations were made on the susceptibility of the olfactory pits but
in the frog Bellamy noted that they are regions of high susceptibility. In gen-
eral it may be said of the sense organs of the head, that the eye is the most
susceptible, the ear vesicles next, and the olfactory pits last. It is therefore
possible to obtain defective eyes in embryos otherwise fairly normal but de-
fective ear vesicles and approximated olfactory pits occur only in embryos
otherwise considerably abnormal. As the matter is not discussed in the text
a word may be said here about the cerebellum. The high susceptibility of the

62 LIBBIE H. HYMAN.

further discussion that these parts of the embryo, shown by me
to be the most susceptible to toxic agents, are also the ones show-
ing the most defective development in the experiments which have
been quoted. In all of these experiments it is evident that the
agents used are inhibiting or depressing agents because as stated
by the authors the development of the eggs subjected to them is
slower than that of the control. Under such depressing condi-
tions the parts with the highest susceptibility or, in other words,
highest metabolic rate, will be inhibited while other parts develop ;
and this is actually the fact. On the other hand, if the circum-
stances permit, such parts can recover more readily than others,
and these same parts may be found developed while other parts
have succumbed. This explains the development of small parts
of the embryo described by Kellicott, Werber and others — iso-
lated eyes, hearts, fragments of brain, etc.

The susceptibility gradients therefore furnish a basis for the
explanation of teratological development. No other conception
which has been advanced does so serve to account for all of the
facts. In particular it seems to me impossible on any other basis
to explain the production of the same terata in eggs fertilized by
injured sperm or by foreign sperm or in cases where the egg is
treated before fertilization as in Gee's experiments. In such
cases a general lowering of the metabolic rate of the egg as shown
by its slower development has occurred and this could produce
specific terata only in case certain parts of the embryo require a
higher metabolic rate for their expression than others.

The application of the metabolic gradient conception to verte-
brate teratology has already been made by several investigators.
Werber ('16) recognized its bearing on the teratological Fundulus
embryos which he produced but failed to grasp the full signifi-
cance of the conception and failed to see that it rendered his own
conception of differential blastolysis superfluous. Newman ('17)

cerebellum is interesting in view of the fact recognized by neurologists that
the cerebellum is a supra-segmental structure added on to the brain stem in
the course of evolution ; and the further fact, discovered by MacArthur and
Jones (':/), that the cerebellum respires about as rapidly as the cerebral
hemispheres, both respiring more rapidly than other parts of the central nerv-
ous system.

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 63

clearly saw that " the principles enunciated by Child serve to
rationalize the results of heterogenic hybridization." He gave
what is for the most part the correct explanation of the terata
originating in his hybridization experiments but fell into a num-
ber of errors because little was then known about the metabolic
gradients in these fish embryos, and he assumed them to be like
those of the flatworms.1 The most complete analysis of verte-
brate teratology which has been made is that of Bellamy ('19) on
the frog, since in this case both the metabolic gradients and the
terata resulting from the differential action of external agents on
the eggs are known.

I have now shown in a general way how the terata produced
experimentally in teleost embryos can be explained on the basis
of the metabolic gradients. Such terata are of two general types,
those due to differential susceptibility, in which the parts of
highest activity are inhibited and defective, and those due to dif-
ferential recovery or acclimation, in which the parts of highest
activity alone survive. A more detailed discussion seems to be
unnecessary in view of the extensive treatment of the matter in
the papers of Child, Newman, and Bellamy already cited. I may,
however, as an illustration of the application of the susceptibility
results to a specific organ take the case of the Fundulus eye. It
happens that in Fundulus, as I have shown, the eyes are very
susceptible with reference to other parts of the body, more so
than in other species of fish. This indicates that the region from
which the eyes arise must be one of very high activity, and as the
data of Stockard show this region must be affected before the
eyes appear in order that defective eyes result. Consequently
inhibition of this region results first in approximated and later
in defective eyes. Now since it happens that in Fundulus this
region is so much more susceptible than in other forms, the oc-
currence of eye defects in Fundulus will also be more common
than in other forms and further it is possible to obtain eye defects

i In particular the statements made by Newman about the gradient of the
heart and circulatory system are quite erroneous. Further the posterior end
of fish embryos is a region of high metabolic rate and embryos with defective
posterior ends are probably due to direct inhibition and are not recovery types
as supposed by Newman.

64 LIBBIE H. HYMAN.

in embryos otherwise nearly normal, if the inhibiting agent is a
rather weak one. This is not possible in other forms ; I venture
to predict that such a result could not be obtained in Tautogo-
labrus but that defective eyes in this species would always be
associated with marked defects of the brain and other parts of
the head. Such is the case in the frog, where cyclopic eyes occur
only in markedly microphthalmic embryos. Owing also to the
high metabolic rate of the Fundulns eye it is possible in this form
for the eye to recover and survive when nearly all other parts of
the embryo are killed. I also venture to predict that the occur-
rence of such isolated and solitary eyes in the absence of other
parts of the embryo will be found to be rather rare and unusual
in other species.1

In conclusion I may reiterate that the study of the metabolic
gradients such as has been made in this paper furnishes a rational
basis for the understanding and interpretation of normal and
teratological development. While the particular organism which
is to develop from a given egg is determined by the hereditary
constitution of that egg, the orderly sequence of development, the
spatial relations and proportions of parts, and the general axial
organization are controlled by physiological, metabolic differ-
ences between different parts of the developing egg. Such physio-
logical differences arise in the final analysis through the action of
external conditions on protoplasm. By modifying in a purely
non-specific, quantitative manner the metabolic differences at dif-
ferent levels, orderly predictable departures from the normal
course of development are obtainable.

VI. SUMMARY.

1. The susceptibility of developing eggs of Fundulns, the din-
ner and the cod to toxic solutions at various stages was studied.

2. In early blastoderms the central region is more susceptible
in Fundulus and the dinner, the peripheral region in, the cod.

1 Monopthalmia, often observed in Fundulns embryos, is simply due to a
greater susceptibility of one side than the other; the eye on the more sus-
ceptible side is inhibited. In the course of my studies on vertebrate embryos,
this asymmetrical susceptibility has frequently been noted although the figures
are drawn as if the susceptibility were always bilaterally symmetrical.

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 65

3. In late blastoderms, the median posterior region of the germ
ring where the embryonic shield is to arise is the most susceptible
region.

4. After the formation of the shield, its anterior portion is the
most susceptible.

5. After the origin of the embryonic axis the anterior end of
the axis is the most susceptible and from this point the suscep-
tibility decreases posteriorly.

6. Sooner or later a secondary region of high susceptibility
arises at the posterior end of the embryo. This secondary region
arises very early in Fundulus, later in the cod, and very late in
the cunner.

7. After the origin of the secondary posterior region, the gen-
eral susceptibility gradient in all three species is a " double " one.
Anterior and posterior ends are the points of highest suscep-
tibility and from them the susceptibility decreases in both direc-
tions towards the middle. Both ectodermal and mesodermal
structures (somites) are involved in the double gradient but the
ectodermal structures (neural tube) are in general much more
susceptible.

8. The heart is highly susceptible (Fundulus). The venous
end of the heart is the most susceptible part of it and from it the
susceptibility decreases towards the arterial end.

9. Besides the general gradients, specific organs may exhibit
high susceptibility. Conspicuous examples of this are the eyes
(especially in Fundulus), the auditory vesicles, and the cere-
bellum.

10. The relations of these gradients to normal development
are considered. It is pointed out that the embryo arises for the
most part from material that does not come from the germ ring
but that later the germ ring contributes to the embryo in degrees
varying in different species. It is further pointed out that the
germ ring type of development is probably a specialization from
a method in which the center of the blastoderm played the chief
role in development. The facts recorded do not support the
theory of concrescence.

11. The oxygen consumption and carbon-dioxide production

66 LIBBIE H. HYMAN.

of developing eggs of Fundulus heteroclitus increase up to the
time when the germ ring is at the equator of the egg. Subse-
quently they decrease relative to the amount of protoplasm but
show an absolute increase owing to the heart beat and other
activity. This period when the respiratory metabolism is greatest
is also the period when the eggs are most readily modified by
external agents.

12. The relation of the susceptibility data to teratological de-
velopment is discussed at considerable length. It is shown that
those parts of the embryo having the highest susceptibility are
those which are most defective in teleost terata and that such
differential susceptibility is therefore the explanation of terato-
logical development. It is also shown that these same parts most
susceptible under extreme conditions may recover if conditions
permit and may develop while the less susceptible parts fail to
recover. Recovery forms of terata are thus just opposite in ap-
pearance to inhibited types.

CITATIONS.

Alvarez, W. C., and Starkweather, E.

'18 The Metabolic Gradient underlying Intestinal Peristalsis. Amer. Jr.

Physiol., Vol. 46, pp. 186-208.
Bellamy, A. W.

'19 Differential Susceptibility as a Basis for Modification and Control of

Early Development in the Frog. BIOL. BULL., Vol. 37, pp. 312-361.
Child, C. M.

'153. Individuality in Organisms. University of Chicago Press.
'i5b Senescence and Rejuvenescence. University of Chicago Press.
'16 Experimental Control and Modification of Larval Development in the
Sea Urchin in Relation to the Axial Gradients. Jr. of Morph., Vol.
28, pp. 65-133.

'17 Differential Susceptibility and Differential Inhibition in the Develop-
ment of Polychete Annelids. Jr. of Morph., Vol. 30, pp. 1-63.
Gee, Wilson.

'16 Effects of Acute Alcoholization on the Germ Cells of Fundulus hetero-
clitus. BIOL. BULL., Vol. 31, pp. 379-406.
Hertwig, P. and G.

'13 Beeinflussung der Mannlichen Keimzellen durch chemische Stoffe. Arch.

f. mikro. Anat., Vol. 83, pp. 267-306.
Hyman, L. H.

'16 An Analysis of the Process of Regeneration in Certain Microdrilous
Oligochaetes. Jr. Exp. Zool., Vol. 20, pp. 99-163.

METABOLIC GRADIENTS OF VERTEBRATE EMBRYOS. 67

Kellicott, W. E.

'16 The Effects of Low Temperature upon the Development of Funduhis.

Amer. Jr. Anat., Vol. 20, pp. 449-483.
Kopsch, F.

'96 Experimentelle Untersuchungen uber den Keimhautrand der Salmoni-

den. Verb. d. Anat. Gesell., icth Versammlung, pp. 113-121.
Loeb, J.

'15 The Blindness of the Cave Fauna and the Artificial Production of Blind
Fish Embryos by Heterogeneous Hybridization and Low Temperatures.
BIOL. BULL., Vol. 29, pp. 50-68.
MacArthur, C. G., and Jones, 0. C.

'17 Some Factors Influencing the Respiration of Ground Nervous Tissue.

Jr. Biol. Chem., Vol. 32, pp. 259-275.
McClendon, J. F.

'i2a An Attempt toward the Physical Chemistry of the Production of

One-eyed Monsters. Amer. Jr. Physiol., Vol. 29, pp. 289-297.
'i2b The Effects of Alkaloids on the Development of Fish (Fundulus)

Eggs. Amer. Jr. Physiol., Vol. 31, pp. 131-141.
Morgan, T. H.

'95 The Formation of the Fish Embryo. Jr. of Morph., Vol. 10, pp. 419—

473-
Newman, H. H.

'08 The Process of Heredity as Exhibited by the Development of Fun-
duhis Hybrids. Jr. Exp. Zool., Vol. 5, pp. 503-563.

'17 On the Production of 'Monsters by Hybridization. BIOL. BULL., Vol.

32, pp. 306-322.
Oppermann, K.

'13 Die Entwicklung von Forelleneiern nach Befruchtung mit radiaumbe-

strahlten Samenfaden. Arch. f. mikro. Anat., Vol. 83, pp. 140-189.
Sumner, F. B.

'03 A Study of Early Fish Development. Experimental and Morphological.

Arch. f. Entw. Mech., Vol. 17, pp. 92-149.
Scott, G., and Kellicott, W. E.

'16 The Consumption of Oxygen during the Development of Fundulus

heterocliius. Anat. Rec., Vol. n, p. 531.
Stockard, C. R.

'06 The Development of Fundulus hetroclitus in Solutions of Lithium
Chlorid with Appendix on its Development in Fresh Water. Jr. Zool.,
Vol. 3, pp. 99-120.

'07 The Influence of External Factors, Chemical and Physical, on the De-
velopment of Fundulus heteroclitus. Jr. Exp. Zool., Vol. 4, pp. 165-201.

'09 The Artificial Production of Cyclopean Monsters : The Magnesium Em-
bryo. Jr. Exp. Zool., Vol. 6, pp. 285-337.

'10 The Influence of Alcohol and other Anaesthetics on Embryonic Devel-
opment. Amer. Jr. Anat., Vol. 10, pp. 369-392.
Werber, E. I.

'16 Experimental Studies on the Origin of Monsters. I. An Etiology and
an Analysis of the Morphogenesis of Monsters. Jr. Exp. Zool., Vol. 21,
pp. 485-583-

68 LIBBIE H. HYMAN.

EXPLANATION OF PLATES.

PLATE I.

FIGS, i and 2. Two stages in the disintegration of an early blastoderm of
Tautogolabrus, showing greater susceptibility of the central cells.

FIGS. 3 and 4. Two stages in the disintegration of an early blastoderm of
Fundulus, showing rupture and disintegration of the central region.

FIG. 5. Disintegration of an early blastoderm of the cod, showing greater
susceptibility of the margin.

FIGS. 6 and 7. Disintegration of a later blastoderm of the cod, showing
greater susceptibility of one region of the circumference and spread of disin-
tegration in both directions from this region.

FIG. 8. A later blastoderm of Tautogolabrus-, posterior half of the blasto-
derm most susceptible.

FIGS. 9 to ii. Disintegration of a late blastoderm of Tautogolabrus. Fig.
9, the normal blastoderm; Fig. 10, bulging of the blastoderm up from the yolk
and disintegration of the central posterior region; Fig. n, further course of
disintegration.

FIG. 12. First appearance of the germinal shield in the cod.

FIGS 13 and 14. Disintegration of a stage like Fig. 12. Fig. 13, disinte-
gration of the shield; Fig. 14, spread of the disintegration around the germ
ring.

FIG. 15. First stage in the disintegration of a later stage of the embryonic
shield of the cod; disintegration beginning at the anterior end of the shield.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE I.

L1BBIE H. HYMAN

7O LIBBIE H. HYMAN.

PLATE II.

FIG. 1 6. Further course of the disintegration shown in Fig. 15.

FIG. 17. Stage of the early embryonic axis in Tautogolabrus.

FIG. 18. Same blastoderm as in Fig. 17, drawn without the yolk; it is
much shrunken. The embryonic region is disintegrating.

FIG. 19. Three stages in the disintegration of the earliest observed em-
bryo of Fundulus. Anterior end to the left. Disintegration begins at the
posterior end, then the anterior end, and proceeds in both directions to the
middle.

FIG. 20. Earliest appearance of the embryo in the cod. Disintegration
beginning at the anterior end of the embryonic shield.

FIG. 21. Further course of the disintegration shown in Fig. 20.

FIG. 22. Later embryo of the cod ; disintegration beginning at the anterior
end of the shield.

FIGS. 23-26. Further course of the disintegration shown beginning in
Fig. 22.

FIG. 27. An early embryo of Tautogolabrus. Disintegration is beginning
at the anterior end of the head.

FIG. 28. Same embryo as Fig. 27, enlarged, showing course of the disinte-
gration along the neural tube.

FIG. 29. An embryo of Tautogolabrus shortly before the closure of the
germ ring.

FIGS. 30-32. Three stages in the disintegration of the embryo shown in
Fig. 29.

FIG. 33. Normal embryo of Tautogolabrus after the closure of the germ
ring.

FIG. 34. Three stages in the disintegration of the embryo shown in Fig.
33. The somites are omitted in the first two drawings.

FIG. 35. An embryo of Fundulus after the appearance of the optic vesicles
and three stages in its disintegration.

FIG. 36. A later embryo of Fundulus and four stages in its disintegration.
The somites are omitted from the latter. The neural tube is characteristically
curved.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE II.

JS

36

tIBBIE H. HYMAN.

72 LIBBIE H. HYMAN.

PLATE III.

FIG. 37. Four stages in the disintegration of a later embryo of Fundultts.

FIG. 38. Three stages in the disintegration of an embryo of the cod after
the formation of the optic vesicles.

FIG. 39. Four stages in the disintegration of an embryo of the cod shortly
before the closure of the germ ring.

FIG. 40. Four stages in the disintegration of an embryo of the cod after
the closure of the germ ring.

BIOLOGICAL BULLETIN, VOL XL.

P1ATE III.

38

f

&

M i

l'-:f-/ 3S»" «*jF

|^ ^

J9

LIBBIE H. HYMAN.

Vol. XL. February, 1921. No. 2.

BIOLOGICAL BULLETIN

AUTHOR S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 7

OBSERVATIONS ON THE DISTRIBUTION AND

HABITS OF THE BLIND TEXAN CAVE

SALAMANDER, TYPHLOMOLGE

RATHBUNI.

EDUARD UHLENHUTH, PH.D.,
ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH, NEW YORK.

When in 1895 the artesian well was drilled at the U. S. Fish
Hatchery in San Marcos, Texas, the first specimens known to
biologists of the blind cave salamander, Typhlomolge rathbuni,
were brought up with the waters from the depths of the ground.
The animals were described by Prof. L. Stejneger. For several
years after this a relatively large number of the blind salamanders,
about 100 a year, were found in the basin of the well, but gradu-
ally the number decreased and lately has been reduced to a few
specimens a year.

When the question arose of subjecting this animal to certain
experiments on metamorphosis, it became evident that a number
of specimens sufficiently large for this purpose could be obtained
only through an extensive search in the actual habitat of the
Typhlomolge. With the aid of a special grant from the Rocke-
feller Institute for Medical Research an extensive study of the
caves of San Marcos and environment was made by the writer
and Mr. C. A. Campbell, at that time instructor in biology at
Coronal Institute in San Marcos, during the months of August
and September, 1916. So far as the number of animals obtained
is concerned, the result was disappointing. But, on 'the other
hand, several observations were made which seem to be of in-
terest as regards the distribution and habits of the Typhlomolge

73

74 EDUAKD UHLENHUTH.

and which furnish valuable suggestions as to the methods which
must be employed in order to procure a large number of animals.
The writer hopes to stimulate a search for these salamanders on
a large scale, in order to make this interesting form accessible to
the experimental biologist who is in need of just such an animal
as Typhlomolge rathbuni for attacking many important problems.

GENERAL CHARACTERS OF THE REGION.

As pointed out, the specimens described first by Stejneger were
found in the basin of the Artesian Well in the Fish Hatchery in
San Marcos, and were carried up into this basin by the flowing
water of the Artesian Well. During a two months' stay in San
Marcos, we secured only two specimens from this basin, but five
other specimens were found in three other localities, i.e., in Frank
Johnson's Well, in Ezell's Cave and in Beaver Cave.

In order to understand the conditions which might have led to
the present distribution of Typhlomolge and because these con-
ditions in the future may be an important guide in tracing the
subterranean channels which the animals inhabit, a careful study
was undertaken. It was found that the conditions in the three
places where we found Typhlomolge are essentially similar to
those existing in the locality from which the water of the San
Marcos Artesian Well is derived.

Before describing the well and the caves in which we found
Typhlomolge it is necessary to point out the geologic peculiarities
of this area of Texas, since these conditions not only led to the
formation of the caves but also to the present distribution of the
Typhlomolge. Whether or not they are also responsible for the
peculiar characteristics of the animal as Eigenmann and Stej-
neger assume, is an important question, the answer to which,
however, cannot be given before extensive experiments on this
species have been carried out.

San Marcos is located on the so-called Balcones scarp line.
This line runs from Austin to Del Rio in a south-westerly direc-
tion and separates in a most distinct way the Edwards Plateau
(north of the line) from the Rio Grande Plain (south of the
line). It forms the escarpments of the plateau towards the plains.
Along this line a faulting has taken place in Eocene time (Hill

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 75

and Vaughan, p. 260), during which the part that now forms the
plain was thrown down and the northern part which now consti-
tutes the plateau was left behind. In consequence of this fault-
ing, any particular geologic stratum now lies deeper on the side
thrown down than on the plateau.

It was apparently this faulting which has led to the formation
of many cracks in the rock layers. The caves near the escarp-
ments of the Edwards Plateau represent gigantic cracks. Be-
sides this factor there is still another cause leading to the forma-
tion of caves in this region. The entire area of the Edwards
Plateau constitutes a huge outcrop of the Cretaceous. In the
soft strata of the various cretaceous formations of the plateau,
numerous caves have been formed by the mechanical force of the
\vater combined with its dissolving action. By this process most
of the rivers of the Edwards Plateau have disappeared almost
entirely from the surface, and their former beds are dry. These
rivers have sunken beneath the surface where they flow in sub-
terranean channels.

THE SAN MARCOS ARTESIAN WELL.

When the Artesian Well of the U. S. Fish Hatchery in San
Marcos (Fig. i) was drilled in 1895, a number of water reser-
voirs were opened up by the drill. At present only the water is
used which rises from a depth of approximately 190 feet. Here
a cave filled with water was opened up ; in it the Typhlomolge
lived. The water in this cave must have been under a pressure
sufficiently high to carry it up 190 feet. The Typhlomolge, thus,
lived most abundantly in water under high pressure and without
any access to air except that present in the water. The water of
this cave belongs to the so-called " sweet water " horizon of the
Edwards limestone in which formation the cave is located.

We measured the temperature of the water as it comes out of
the tube of the well as approximately 21.5° C. Among the fauna
of the cave from which the water of the San Marcos Well rises,
are particularly conspicuous two crustaceans, both unpigmented
and eyeless, an isopod, Cirolanides texensis and a decapod, Palce-
monetes antrorum. The latter species is of particular importance,

76

EDUARD UHLENHUTH.

since so far it has been found to occur in all localities which are
inhabited by the blind salamanders.

The cave of the San Marcos Well, thus, is characterized in the
following manner: (i) It is situated in the Edwards limestone.
(2) It contains water derived from the " sweet water " horizon.

FIG. i. Basin of the Artesian Well of the U. S. Fish Hatchery in San Marcos.

(3) The temperature of the water is approximately 21.5° C. (4)
The water is inhabited by the decapod, Palccmonctcs antrorum.

FRANK JOHNSON'S WELL.

Approximately two miles southwest of the San Marcos court
house (see map, Fig. 2), the flat valley of the dry Purgatory
Creek crosses the Balcones scarp line opening here into the flat
valley of the San Marcos River. Its northern slopes are formed
here by the San Marcos Hill. Purgatory Creek originates near
the Devil's Backbone, the divide between the Guadalupe and
Blanco Rivers, at Boyett's Farm, about 14 miles northwest of
San Marcos. It is dry at present, but several of the older in-

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI.

77

habitants claim that this creek had running water in it until about
50 years ago. At present only a few water holes are left in the
upper course of the valley and several sink holes have formed in
its lower course. These are filled temporarily with rain water.
In time of severe cloud bursts the water in the creek becomes a

»* II /''**-.

&•) ^"vi^- -»—'"''

\ r t»f

FIG. 2. Map of San Marcos Area.

1. San Marcos. 5. Ezell's Cave.

2. Artesian Well of U. S. Fish 6. Frank Johnson's House.

Hatchery. 7. Frank Johnson's Well.

3. Head of San Marcos River. 8. Swift's Cave.

4. Beaver Cave.

torrent rising to a height of 8 feet, but it disappears completely
within several hours.1 Purgatory Creek has now become a sub-
terranean creek. Mr. Frank Johnson's farm is located near
where the creek crosses the fault line.

i The general character of a creek like this may be found described in Hill
and Vaughan, page 207.

78 EDUARD UHLENHUTH.

Mr. Johnson informed me soon after my arrival in San Marcos
that the blind white salamander has been seen in his well, and in
fact this well has yielded us more salamanders than any other
place. It is shown in Fig. 3.

The well is located in the valley of Purgatory Creek, a short
distance above where the creek enters the plain. Part of the flat
valley is visible in the figure. Near the well is a sink hole (Dris-
kel's Water Hole, see diagram, Fig. 4). The well was dug from

FIG. 3. Frank Johnson's Well. Showing the well house and at the right of
the well house the dry and flat valley of the Purgatory Creek.

a level of 613 feet1 above sea (San Marcos Court House 620
feet) to a depth of 317^ feet. There a cave was struck which
now communicates with the well through a slit in the well bot-
tom, as indicated in the diagram (Fig. 4). From this slit the
water rose to from 3 to 5 feet in the well. This makes the sur-

1 Altitudes above sea level were measured by means of an anaeroid barome-
ter and therefore are only approximately correct (within several feet). Di-
mensions other than altitudes were measured directly, except when otherwise
stated.

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI.

79

face altitude of the water about 584 feet. Mr. Johnson claims
that the water is flowing. There is no doubt that Johnson's Well
communicates with the subterranean Purgatory Creek. As in
the case of the San Marcos Artesian Well, the water in this com-
pletely water filled cave must have been under a pressure suffi-
ciently high to lift it to 3 feet in the well. It again is evident that
the Typhlomolge prefer to live in water under high pressure and
in caves which are filled entirely with water. The water of John-
son's Well has the same temperature as that of the San Marcos
Artesian Well and also has the same taste. Besides the Typhlo-
molge, Frank Johnson's Well contains also the Palcemonetes an-
troruin and the Cirolanides texensis.

\j4vwi

ov s

FIG. 4. Purgatory Creek Valley and Frank Johnson's Well. Diagrammatic
section reconstructed from several cross sections. The figures indicate alti-
tude above sea level in feet.

Thus, though the water of the Frank Johnson Well represents
the subterranean Purgatory Creek, it shows great similarity to the
water of the San Marcos Artesian Well. Particularly the pres-
ence in Purgatory Creek of 3 species typical of the artesian well
would suggest that in some way the Purgatory Creek water is in
communication with the so-called sweet water horizon near the
San Marcos Artesian Well.

In Frank Johnson's Well the Typhlomolge were seen to pass
through the slit from the cave into the well. During my stay in
San Marcos, the water in the well was too high to catch the sala-
manders directly and for this reason traps were submerged in the
well. These were ordinary minnow traps. In the beginning they

8O EDUARD UHLENHUTH.

were supplied with various kinds of bait, but in this way only the
crustaceans mentioned above were caught. The Typhlomolge
did not seem to react to the bait, and later on when I observed the
animals in the laboratory, it became evident that the instinct of
hunger is not sufficiently strong in the Typhlomolge to make
them go into traps ; it is in fact very difficult to make these ani-
mals eat. Later on the traps were placed with one opening di-
rectly in the slit ; animals passing out from the slit had to go
directly into the trap. In this way 2 Typhlomolge were caught in
Johnson's Well, one in August, 1916, and another in September,
1916. After I had left, n more Typhlomolge were found by Mr.
C. A. Campbell and Mr. Rufus Smith who from time to time
looked after my traps. Thus, Frank Johnson's Well yielded us
13 specimens of Typhlomolge. They were caught as shown in
the following table. The number is, however, too small to war-
rant any conclusions as to a possible influence of the season upon
the frequency of the occurrence of Typhlomolge.

August, 1916 i

September, 1916 i

November, 1916 2

December, 1916 i

January, 1917 2

April, 1917 i

Summer, 1917 2

November, 1917 3

One of the greatest difficulties encountered was to find a
method of shipping the animals from San Marcos to New York ;
most of them did not survive the trip. In fact, only two ever
reached the laboratory alive. The first seven specimens caught
were taken on the train in a bucket filled with water. The
jarring killed six. Among the eleven caught later on, only one
survived the trip. Its safe transfer was accomplished by a fortu-
nate incident. The animal was packed in a fruit preserving jar
filled entirely with water and shipped in the winter. On arrival
it was frozen tightly in a block of ice. This animal survived for
one year in the laboratory. The only thing it could be made to
eat were newly hatched larvae of Ambystoma maculatum. Though
kept for most of the time in a dark room, the skin which in the

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 8 1

beginning was white with a bluish, mother-of-pearl gleam, had
darkened somewhat.

It should be pointed out here that slow reaction to food as
exhibited by the Typhlomolge'*- is noteworthy in regard to certain
findings of Miss E. T. Emmerson, who claims, upon anatomical
reasons, a close relationship between Typhlotnolge and the larvae
of Eurycea rubra. We are keeping a large number of such
larvae in the laboratory and contrary to my experience with the
larvae of Ambystoma and other salamander larvae, these larvae
react very slowly to food. In fact, it is impossible to make them
eat every day aside from the fact that most of the individuals of
this species will eat only at night.

EZELL'S CAVE.

Ezell's Cave was opened up several years before the San
Marcos Well was drilled. The entrance to the cave is located
on the southwest slope of the San Marcos Hill (see map, Fig. 2),
wrhere it slopes down to the valley of Purgatory Creek about 2
miles W.S.W. of the San Marcos Court House, and not far
from a little ravine, the bed of the dry City Boundary Creek, a
tributary to Purgatory Creek. This location of Ezell's Cave indi-
cates that it belongs to the Purgatory Creek System, the river
found in it probably being the subterranean course of the City
Boundary Creek.

Ezell's Cave distinctly exhibits the aspect of a large crack in
the strata of the hill, brought about by dislocation of the strata
towards the Purgatory Creek Valley. The entrance to the cave
(approximately 6/0 feet above sea level) is part of a 62 ft.
slit in the surface (Fig. 5), which for the most part is closed
up by large rocks and runs from N.N.W. to S.S.E., that being
the direction of the long axis of all the various parts of the cave.
As the diagrammatic cross and longitudinal sections (Fig. 6 and
7) indicate, the entire slit so far as accessible is divided into two
compartments by means of the rock masses which were thrown
down during the process of dislocation and following corrosion.

i Normann, who kept a specimen of Typhlomolge in captivity, also reports
great difficulty in making the animal eat.

82

EDUARD UHLENHUTH.

These masses of debris form the bottom of the first story and in
the N.N.W. corner leave open a small hole ("entrance hole"),
2^4 feet wide through which a narrow canal ("tube") may be
reached which after running along the main axis of the slit for
a short distance leads down into the second story or water room.

FIG. 5. Entrance to Ezell's Cave.

This compartment of the cave contains a large body of water
(Fig. 8).

This pond is not formed by water which drains through the
strata forming the roof of the cave nor by water flowing directly
into the entrance of the cave, as the slope of the hill is drained in

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI.

the course of rain. The pond is formed by a subterranean river,
which is evident from the fact that the water is flowing, though
hardly in a perceptible manner. The flow can be observed from
the dislocation of bodies dropped into the water at the N.N.W.
end of the pond. If the water is not disturbed such bodies will
arrive, in the course of an hour or so, at the S.S.E. end, thus

By means of a collapsible

indicating the direction of the flow.

RUttu.de

10

FIG. 6. Ezell's Cave. Diagrammatic section reconstructed from several

cross-sections.

boat which was brought down into the water it is possible to
follow the course of the subterranean creek towards N.N.W.,
(Fig. 9) for a distance of about 91^ feet. The crack extends,
however, beyond this point and by climbing over a number of
rocks the creek can be seen to continue in this crack. But we
had no opportunity so far to explore this part of the cave.

The greatest depth of the water is 13^2 feet, as far as it can

84

EDUARD UHLENHUTH.

be measured. It is, however, not possible to ascertain exactly
the depth of the water and of the crack, since the water is
covered in part by the overlapping wall of the crack forming
a ledge over the water (diagram Fig. 6 and photograph Fig. 9).
Underneath this ledge the ground can be seen to slope down very
deeply ; it is possible by means of a strong light to see a f unnel-

RUitude

N-S.W

M a in -H oo TIX - --/--

Entrance HoXe
lute

10

FIG. 7. Ezell's Cave. Diagrammatic section reconstructed from several longi-

tudinal sections.

shaped crater opening at the deepest part of the lake in which
no bottom can be seen.

The entire crack, with the water which it contains, is located
in the Edwards limestone ; but as pointed out above, the structure
of the cave would indicate that this crack may extend into the
deeper lying strata.

The distance from the entrance down to the water surface is
94 feet, which makes the level of the water about 577 feet. The
altitude above sea level of the entrance of the cave was measured
merely by means of a barometer, but the figure approaches the
altitude of the water surface in Frank Johnson's Well near

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 85

enough ; the water levels in Frank Johnson's Well and in Ezell's
Cave are approximately equally high.

The water is of an extreme clearness and of bluish color,
typical also of the water of the sweet water system. It also
tastes like this water and has the same temperature (21.5°).
Using a sufficiently strong light one discovers immediately a
great number of Palccnionetes antrorum swimming near the sur-

, .

FIG. 8. Water-room in Ezell's Cave.

face of the water, which thus contains also the same species of
animals as were found in the water of the San Marcos Artesian
Well and in Frank Johnson's Well.

Hence the water in Frank Johnson's Well and that in Ezell's
Cave have a number of characteristics in common. They have
the same taste, are of the same temperature, and their levels are
equally high. They harbor the same species of animals. From
their characteristics and from their location it seems that they
are parts of the subterranean Purgatory Creek System.

Furthermore, both of these water bodies have certain most

86

EDUARD UHLENHUTH.

conspicuous characteristics in common with the water of the
San Marcos Artesian Well. They are of the same temperature
and contain the same fauna. One naturally would think of a
direct communication between the Purgatory Creek System
and the caves which supply the San Marcos Artesian Well.

FIG. 9. Ezell's Cave Lake. Showing the overlapping ledge.

«

We caught only one animal (78.5 mm.) in Ezell's Cave. It
was sitting quietly near the bank of the river where the water is
shallow, and did not seem to mind pebbles dropped down into the
water near it, nor the glare of the light from two Columbia dry
cells. We spent 12 days in the cave under the most varied con-

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 87

ditions, and conducted a most extensive search for Typhlomolgc.
Hence the scarcity of this species is somewhat perplexing. It is
possible that the animals prefer to stay further down in the
passages and cracks filled completely with water under high pres-
sure, an assumption which is supported by the circumstances
under which the animals were found in both the artesian well and
Frank Johnson's Well. It may be that they rarely and only by
some incidental circumstances are induced to come to the more
open bodies of water.

So far as is known to the writer, the specimen of Typhlomolge
caught in Ezell's Cave in 1916 is the first and only one positively
known as having come from this locality. But it is claimed by
people in San Marcos, as Mr. S. N. Stanfield, teacher of biology
in the Texas Normal School in San Marcos, informed me, that
the first two Typhlomolge ever seen were found in Ezell's Cave,
i Y-2. years before the well was drilled, in a small boat, which
had sunk in Ezell's Cave Lake.

BEAVER CAVE.

Not far from the entrance of Ezell's Cave on the southwest
slope of San Marcos Hill and at an altitude of 652 feet above sea
level, near the dry bed of the City Boundary Creek is situated the
entrance to Beaver or Wonder Cave. The location of the cave
would indicate that it belongs, like Ezell's Cave, to the Purgatory
Creek System.

Beaver Cave represents the aspect of a straight running crack
in the strata of the Edwards limestone, the same as Ezell's Cave ;
this crack, in part, has been widened out and its walls have been
smoothed down by the action of the water (Fig. 10). Its bottom
is made up of huge masses of broken-down rocks which form, at
some places, high cliffs and rock masses, dividing the entire cave
horizontally in a number of rooms connected by narrower tubes
with one another, and vertically into several compartments. Fig.
II represents a diagrammatic longitudinal section through the
cave, which gives an idea of the construction of this cave. In
Fig. 10, which was taken parallel to the longitudinal axis, the slit-
like shape of the cave is shown ; it can also be seen how smooth

88

EDUARD UHLENHUTH.

the walls have been washed by the water entering easily through
the thin roof of the cave.

The longitudinal axis of Beaver Cave runs from N.N.E. to
S.S.W,, forming an angle of approximately 25° with the longi-

FIG. io. Interior of Beaver Cave. Photograph taken from rock 34 towards
board rock. In back of the right hand side wall at its lower end, the opening
of the " tube " is visible.

tudinal axis of Ezell's Cave ; the length of the entire slit is nearly
500 feet. It is claimed that there is a direct connection between
Beaver Cave and Ezell's Cave. We could not verify this state-
ment, and it seems certain no one has actually found a connec-
tion. We found that at .v in room VI. (see Fig. n) a number of

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI.

89

tightly packed rocks and masses of
gravel make further penetration im-
possible at present and that the lo-
cation of both caves and the direc-
tion of their main axes are not in
favor of the statement mentioned
above.

The deepest depression in the
bottom of Beaver Cave is found in
the room indicated in the diagram
Fig. nas" Weil-Room." The bot-
tom of this floor is 62 feet below
the surface and therefore at a level
of 590 feet above sea level. As seen
from the height of the water level
in Johnson's Well and in Ezell's
Cave, no water of the Purgatory
Creek System should be present in
Beaver Cave. And in fact when
the cave was discovered there was
no water found. But a well drill-
ing made at that time from the sur-
face above the Well Room had in-
dicated the presence of water only
a few feet beneath the bottom of
the Well Room. Therefore, a hole
was dug in the bottom of the Well
Room which led to water at a depth
of about 3 feet or at the same level
as the surface of the water in Ezell's
Cave and Frank Johnson's Well
(see Fig. 13).

At present one finds in the Well
Room of Beaver Cave a rectangular
basin approximately 6 feeet in
length, 3 feet in width and 6 feet in
depth, the bottom of which is cov-
ered with mud and rocks, and the
walls of which are lined with logs.

\

' O

9O EDUARD UHLENHUTH.

This basin is filled with water half of its depth. Hence the surface
of the water stands at the same level with the surface of the water
in Frank Johnson's Well, and the suggestion seems justified that in
this basin again part of the Purgatory Creek System was opened
up. The water has the same taste as the water of the other
localities mentioned and also has the same temperature (21.5° C.).
In which way, however, this basin in Beaver Cave could be con-
nected with the other localities cannot be stated with certainty at
present, since the log lining of the wall made it impossible to
search more closely whether or not the rocks of the wall contain
any larger cracks or crevices. It also was not determined
whether the water is flowing. But its clearness and the fact that
the mud when stirred up disappears in a relatively short time
would speak in favor of a slight current in the water. There is,
however, one fact which hardly could be explained in any other
way than that the water in the basin must be in connection at
least at certain times with some larger bodies of water. The
well in Beaver Cave contains both the Palceinonetes antronnn and
the Cirolanides texensis, animals the transmission of which to the
basin since it was constructed must have taken place by means of
water currents which drive water from certain water bodies
(harboring these animals) through the well.

Hence it is most probable that the water of Frank Johnson's
Well, of Ezell's Cave and of Beaver Cave is the water of the sub-
terranean Purgatory Creek System.

In the well of Beaver Cave two Typhlomolge were caught, one
by means of a dip-net, the other in a trap which was laid with its
opening just in front of a hole into which the animal had been
seen to pass. One specimen was 82 mm. in length, the other
one the largest caught measured 120 mm. Both these animals
were observed for some time before they were actually caught ;
they proceeded to move in characteristic fashion — as described
very accurately by Normann — by intermittent walking and rest-
ing in the presence of light. Even when the rays fell directly
upon them, they did not seem to be disturbed. In this respect
our observations made in the animal's natural habitats, agree
verv well with the observations made bv Normann in the labora-

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 91

tory. Pebbles and a pocket-knife dropped into the water near
the animal did not change its behavior; we have not found that
the Typhlomolge as Normann claims possesses a specially high
sensitivity towards disturbances of the water. Once stirred up
the animals immediately swim towards the walls, and if they
cannot find cover immediately, they swim along the wall toward
the surface pushing out their snouts above the surface.

Before I was acquainted well enough with the general situa-
tion in the localities in question and before I had other facts
indicating a possible connection between Beaver Cave and the
Purgatory Creek, the occurrence of the Typhlomolge in the
Beaver Cave well was puzzling, since it seemed to be difficult to
explain how they could have been transferred to the well. In an
anatomical study performed on Typhlomolge rathbuni, E. T.
Emmerson points out the close relationship existing between
Typhlomolge and Eurycea (Spelerpes}, in particular Eurycea
rubra and suggests that Typhlomolge may be the larva of an
unknown species of the genus Eurycea. The writer of this article
has a large number of larvae of Eurycea rubra under observation
and finds that in certain habits (feeding and especially the push-
ing out of the snout above the water when aroused) a remarkable
resemblance exists between Typhlomolge and Eurycea rubra, a
resemblance which was not observed by the writer in larvae of the
many other species of salamander closely watched in the labora-
tory. Concerning, however, the assumption that Typhlomolge
is the larva of some species of Eurycea, this meets with one diffi-
culty if it should mean that this species is still in existence.
Ezell's Cave and especially Beaver Cave were closely searched
for the presence of other salamanders. None were found in
Ezell's Cave. In Beaver Cave, however, Mr. Campbell found
about 20 specimens all belonging to the species Plethodon gluti-
nosus; this is the only salamander which we could detect in these
and other caves of the area around San Marcos. In view of this
fact it appears that the suggestion as to whether or not Typhlo-
molge is the larva of a species represented at the present time
also by metamorphosed specimens would be hardly more than
speculation. It is, however, certain that it would be of the

92 EDUARD UHLENHUTH.

greatest value to raise the Typhlomolge, in order to study closely
their mode of propagation, development and to subject these
animals to certain experiments indicated by our present technic
in the study of the metamorphosis of other salamanders.

In connection with the metamorphosis of Typhlomolge it may
be pointed out that Miss Emmerson has made a statement which
is so important that it arouses curiosity as to why it has attracted
so little attention. Miss Emmerson searching for the organs of
internal secretion of Typhlomolge found that the animal possesses
a thymus gland, but she could not find a thyroid gland. If the
lack of a thyroid gland could be confirmed — and we are prepar-
ing some of our specimens for examination with that end in
view — Miss Emmerson's discovery will explain why the Typhlo-
molge cannot metamorphose at present, since Allen has demon-
strated that larvae of frogs and toads whose thyroids were ex-
tirpated did not metamorphose, though the controls with intact
thyroids all metamorphosed. Do the Proteidas (Proteus) possess
thyroids, is the lack of the gland common to all of them? And
what are the reasons for the atrophy of the gland? These are
problems which call urgently for investigation.

From the facts mentioned above it is certain that the Typhlo-
molge inhabit the subterranean waters which constitute the
Purgatory Creek System and a subterranean water channel which
supplies the San Marcos Artesian Well. These two systems are
located north and south respectively from the Balcones scarp
line. On account of the faulting, though both the Purgatory
Creek Caves and the Artesian Well Cave are located in the same
geological formation, the latter cave occupies a position several
hundred feet deeper than the Purgatory Creek Cave; this is
indicated in the diagram, Fig. 12. The water in both systems is
of different origin, as may be seen from this diagram. The
water of the Artesian Well is the so-called "sweet water," which
on the plateau, i.e., in the region of the Purgatory Creek System,
is carried in beds below those in which the caves of the Purga-
tory Creek System are located. The " sweet water " is caught by
the basement beds of the Cretaceous, the Travis Peak and Glen
Rose formation, where they outcrop on the plateau, and is carried

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI.

93

down along the slanting stratum beneath the geologically higher
situated Edwards limestone and towards the fault. Along the
fault, however, the continuity of the water-bearing strata is
broken and they come to lie in one level with the Edwards lime-
stone of the plain ; thus, here the water is forced from the Glen
Rose formation into the Edwards limestone.1

fel°-~f£

H<H^t (blQ\

abo—

V'VVY\£, -

»oo

FIG. 12. Ezell's Cave. Balcones scarp line and Artesian Well of U. S.
Fish Hatchery in San Marcos, Texas. Diagrammatic section showing the
position of northern and southern part of the various cretaceous formations
to each other after the dislocation of the Rio Grande Plain in Eocene time
accomplished by faulting. The figures on the left-hand side of the diagram
indicate altitude above sea level in feet.

The water of Johnson's Well, Ezell's Cave and Beaver Cave,
however, is the river water of the subterranean Purgatory Creek.
The level of the Purgatory Creek in these localities at present is
at an altitude of about 580 feet, that of the Artesian Well 360
feet.

Hill and Vaughan, page 315.

94

EDUARD UHLENHUTH.

But as indicated above, it is quite probable that a direct com-
munication has been established between these two systems by
means of channels which according to our calculations would have
a depth of about 200 feet ; if this is a fact, the relation between
the various bodies of water in question would be as shown in
the diagram of Fig. 13.

FIG. 13. Purgatory Creek System. Artesian Well and San Marcos Springs
at San Marcos, Texas. Diagrammatic section reconstructed from several
sections, showing the way in which these waters probably communicate with
each other. The figures indicate altitude above sea in feet.

It is of great importance to ascertain whether or not such a
communication exists, since this would facilitate following the
Typhlomolge along the course of travel and since it would permit
conclusions as to the mode of the distribution of the species.
Besides the suggestive structure of Ezell's Cave there are a num-
ber of facts which are in favor of the existence of a communica-
tion. If no connection between the two systems exists it would
mean that the Typhlomolge lived in the subterranean rivers be-

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 95

fore the present southern and northern parts of the Edwards
limestone were separated from each other, and that after the
dislocation in Eocene time part of the species was caught in the
caves of the Edwards limestone of the San Marcos area south
of the Balcones where it lived completely isolated from the rest of
the species. Since the specimens obtained from Ezell's Cave and
the Artesian Well are identical, it would mean either that the
species remained absolutely unchanged since Eocene time, or if
it changed, underwent exactly similar changes in the open ponds
of the subterranean Purgatory Creek and in the completely closed
and water-filled subterranean caves of the Artesian Well. It is
evident that none of these possibilities is probable.

Not only the Artesian Well at San Marcos but numerous other
artesaan wells along the Balcones escarpments are supplied from
the sweet water horizon ; yet from none of them, except the San
Marcos Well, Typhlomolge has ever been reported. This would
be explained if the San Marcos Well contains besides the sweet
water also the Purgatory Creek water, since this certainly could
not be true for the other wells. Probably the Purgatory Creek is
the original habitat of the Typhlomolge and later on the animals
migrated down to the water channels of the Artesian Well.

Also in none of the fissure springs of the Balcones scarp line,
not even in the San Marcos springs though they all come from
the sweet water reservoirs, Typhlomolge ever has been collected.
The same explanation as to the artesian wells could be applied
to these springs, if a communication exists between the Purga-
tory Creek System and San Marcos well.

Finally an incident may be mentioned here which also would
speak in favor of the existence of a direct communication between
the Artesian Well and the Purgatory Creek System. Mr. Mark
Riley, superintendent of the U. S. Fish Hatchery, informed me
that in the basin of the Artesian Well a number of catfish were
kept at one time, but they disappeared gradually from the basin
and it is claimed that they migrated into the tube of the artesian
well. The writer is not prepared to form an opinion concerning
the probability of such migration. One day, however, while I
was looking for Typhlomolge in Ezell's Cave, I saw some fishes

96 EDUARD UHLENHUTH.

hiding behind the rocks. Shortly after this we caught two fishes
by means of hooks which were placed near the rocks where I had
seen the fishes ; both were catfish. And they were the only speci-
mens of fish which I ever saw in Ezell's Cave during the 12 days
I spent there. If these were identical with the individuals kep?
in the basin of the Artesian Well, it certainly would be proof of
the existence of a communication between the Purgatory Creek
System and the San Marcos Artesian Well. It would be of great
importance to trace the course of the water in Ezell's Cave and
Johnson's Well down to the reservoir of the Artesian Well. As
suggested by the possible migration of the catfish, such methods
could be easily designed and will be employed as soon as the in-
vestigations can be continued.

In case of a connection between the two systems, the water
contained in each one would be a mixture of the Purgatory Creek
water and the sweet water. In all four places in question the
water has the same taste and the same temperature. It contains
besides the Typhlonwhjc a number of typical species, among
them the Palcemonetes antrorum, which I found to occur in all
four localities.

OTHER LOCALITIES IN PURGATORY CREEK VALLEY.

After it was found that the Typhlomolge inhabit subterranean
regions probably representing the Purgatory Creek System, it was
interesting to visit other caves of the Purgatory Creek. One of
them is Swift's Cave, on the slope closing the valley towards
southwest and abuot I mile above Frank Johnson's Well. The
entrance to the cave is situated at an altitude of 701 feet above
sea level. Though no water could be reached so far — according
to what has been said above, it would have to be found about 114
feet below the entrance — there is in this cave a narrow tube
leading down which has not been followed ; further examination
may reveal the presence of some passages to the water.

Further up the valley 14 miles above San Marcos, Boyett's
Cave is located at an altitude of about 1,100 feet; there the
Purgatory Creek valley starts. No water was found in Boyett's
Cave down to a depth of 50 feet and no passages leading further

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI.

97

down were discovered. But in the large main hall of this cave
one notices along the walls a whitish deposit for about 3^-2 feet
above the ground and forming a straight horizontal line running
along the walls, as seen in Fig. 14. This indicates the former
presence of water in this cave. Probably with the general dis-

FIG. 14. Boyett's Cave.

appearance of the water from the Purgatory Creek valley and its
fall to deeper levels, flowing water has disappeared from the cave.
There are, however, a number of small shallow pools (several
inches deep) formed from dripping water in the sandy bottom

98 EDUARD UHLENHUTH.

of the main hall. In these pools certain crustaceans are found in
large numbers, which according to Dr. Ortmann, are at least very
closely related to if not identical with the species Stygonectes
flagellatus? an amphipod knowrn from the San Marcos Well.
Thus, this animal, which through its mode of living is well
adapted to the conditions prevailing at present in Boyett's Cave,
is the only remnant there of the Purgatory Creek System fauna.

THE SAN MARCOS SPRINGS.

According to Hill and Vaughan it is quite certain that the San
Marcos Springs, like all other fissure springs along the escarp-
ments of the Edwards Plateau, are of the same origin as the ar-
tesian wells of this area, and hence the water of the .San Marcos
Springs comes from the same reservoir which supplies the Arte-
sian Well at San Marcos Fish Hatchery. We might expect there-
fore that between these two localities ways of communication
exist along which the Typhlomolge may travel.

We have not so far subjected the San Marcos Springs to a
thorough examination, but a brief mention may be made of cer-
tain facts valuable for future exploration. The water of the San
Marcos Springs comes from funnel-like depressions of the sur-
face (see diagram, Fig. n), and forms a little lake which is the
head of the San Marcos River. The openings through which the
water emerges lie deeper (at an altitude of only 532 feet) than
the surface of the water of Purgatory Creek. The surface of
the lake which is artificially dammed, is 559 feet above sea level.
The temperature of the water in a little spring on the bank of the
lake is 21.5° C,, like that of the water of the Purgatory Creek
system and the Artesian Well.

Towards the south the valley at the head of the River continues
and forms the bed of the San Marcos River in which the water
is flowing, but this valley can be traced also north of the springs,
though here it is dry.

Unconfirmed claims have been made that the "white sala-
mander " was seen at the head of the river, but that it had de-
veloped eyes and turned brownish. These statements are, no

i For further information see Benedict and Weckel.

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 99

doubt, due to the occurrence there of the larvse of other sala-
manders which have been mistaken by the layman for Typhlo-
molge. But even if the animal should come up into the lake, it
would be quite difficult to find it there; unaccustomed to such
rapacious enemies as certain fishes which abound in the lake, the
blind Typhlomolge would soon fall a victim.

There are, however, two localities further up in the dry valley
which it might be important to examine. One is a hole resembling
a well hole because of its regularity. It is about 6 feet deep. Mr.
Bidler who is well acquainted with conditions as they were 20 to
30 years ago in this area, informed the writer that at one time this
hole was much deeper and contained a small body of water. He
assured me that one day in the water of the hole two white
Typhlomolge were seen. At any rate this hole should be pre-
pared for examination by removing the gravel and rocks and thus
penetrating to the water, the level of which would be approxi-
mately that of the water in Purgatory Creek. The place could be
easily prepared so as to make trapping there a success.

Near this place on one of the slopes of the valley is situated
another hole (on the property of Mr. Mark Riley), which much
resembles the entrance to Beaver Cave. In former years, before
the water which abundantly drains into that hole had washed
gravel into it, one could penertate it for some distance and reach
a place at which the sound of water could be heard.

The writer believes that the exploration of the two places men-
tioned would lead to a definite knowledge about the presence of
Typhlomolge in the San Marcos River valley.

OTHER PLACES TO BE EXPLORED.

It is clear that Typhlomolge cannot be procured in abundant
number by collecting in one or two places only, since these ani-
mals pass only in small numbers from the deep water-filled caves
into the open water bodies of the higher horizons. Successful
collecting must have as a basis the discovery of a large number of
places where some of these animals can be found. Traps must
be laid in all these places and watched for some time. Assuming
that from three such localities 5 specimens could be obtained in

IOO EDUARD UHLENHUTH.

the course of two months, as was the case in Ezell's Cave,
Beaver Cave and Johnson's Well, a steady and skillful worker
could collect from 18 such places 180 specimens in a year, a num-
ber sufficiently large to start experimental work on the species
and to keep a sufficient number for breeding stock. For this
reason an attempt will be made to mention briefly a number of
other places where Typhlomolge may possibly be found.

There are two caves on the ranch of Mr. Bender at Spring
Branch, 40 miles above San Marcos and about 1,100 feet above
sea level. One is a narrow channel through which the head water
of Spring Branch Creek passes out. The channel is filled almost
to the top with water but it is possible to penetrate it to a depth
of 350 feet. Since the water is flowing quite rapidly, it is not
likely to contain Typhlomolge, but a more thorough search might
be conducted. The other cave represents a narrow crack in the
strata containing water at a depth of 45 feet. It is only a small
pool, which, however, is part of a larger body of water covered
by overlapping ledges. The temperature of the water is 20.5° C.
Besides frogs, some other animals inhabit this pool. They could
not, however, be identified, as upon our approach they imme-
diately dived underneath the ledge.

More important still is the water on Mr. Bremer's ranch, at
the water hole of the Cypress Fork in Hays County, a tributary
branch of the Blanco River, about 1,000 feet above sea level.
The water hole (Jacob's well) itself is filled with blue water
which has a temperature of 22.5° C. On account of the large
black basses inhabiting the hole one would not expect to find
Typhlomolge there. But further up on one of the slopes of the
dry valley is located the entrance to a cave in which the water
(probably of Jacob's well) could be reached. We penetrated to
a place where a number of small holes perforate the bottom of
the cave; pebbles thrown into the holes evoked the sound of
rather deep water. By dislocating a large rock, it would be pos-
sible to make one of the holes large enough to gain access to the
water.

It might be valuable to mention a few places in which, accord-
ing to Mr. S. A. Stanfield, Typhlomolge have been seen :
Burnet Cave, Kendall County, near Burnet.

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. IOI

A spring near Twin Sister Mountain, Hays County, 2 miles from

Wimberly.
A spring near Ozona, 100 miles from San Marcos.

SUMMARY.

1. At present it is certain that Typhlomolge rathbuni inhabits
the subterranean water of the Purgatory Creek System just north
of the Balcones scarp line and one mile further up, and the caves
of the Artesian Well of the U. S. Fish Hatchery at San Marcos,
which seem to be in direct communication with the Purgatory
Creek System by means of channels about 200 feet deep.

2. The populations of the species Typhlomolge rathbuni north
and south of the present Balcones scarp line have not been sepa-
rated from each other by the process of faulting in Eocene time,
but have developed in unrestricted communication with one
another.

3. No certain data are available as regards the occurrence of
Typhlomolge in the San Marcos Springs and in the dry valley of
the San Marcos River north of the Springs. Since .the Springs
come from the same water reservoir as the Artesian Well, further
investigations should be conducted.

4. All the localities containing Typhlomolge are located in the
Edwards limestone region, but the caves of the Artesian Well are
200 feet deeper than the rest.

5. Typhlomolge have been found in the Purgatory Creek Sys-
tem at an elevation of approximately 585 feet above sea level.
Where this level could not be reached as in the upper Purgatory
Creek valley, only remnants of the Purgatory Creek System fauna
(Stygonectes flagellatum) were found.

6. The water inhabited by Typhlomolge seems to be slowly
flowing water.

7. The temperature of the water is approximately 21.5° C.
and it is inhabited by the decapod Palazmonetes antrorum. Since
the latter animal is much more numerous and can be detected
much easier than the Typhlomolge, its presence may be taken as

an indication that the place is promising as regards the presence
of Typhlomolge.

IO2 EDUARD UHLENHUTH.

8. The rarity of the Typhlomolge seems to be due to the ani-
mal's habit of preferring deep lying cracks or crevices, com-
pletely filled with water at a higher pressure than exists in the
more open bodies of water located at higher levels.

9. As regards the habits of the Typhlomolge in its natural
habitat we were able to confirm Normann's observations made in
the laboratory in respect to the peculiar mode of walking of this
animal and its indifferent attitude to light. But we did not find
the animal particularly sensitive to water waves.

10. In feeding and swimming when aroused, Typhlomolge
shows a close resemblance to larvae of Eurycea rubra.

u. The assumption, however, that Typhlomolge is the larva
of some unknown and still existing species of the genus Eurycea
as made by Emmerson could not be confirmed, since with the ex-
ception of the species Plethodon glutinosus no tailed Batrachians
were found in the caves. More important than this assumption
is the fact that Typhlomolge, according to Emmerson, lack a
thyroid, which would explain why these animals cannot meta-
morphose.

12. In order to collect a large number of specimens necessary
for experimental work and intensive study of the species, as
many places as possible must be discovered which may contain
Typhlomolge, and collecting must be conducted simultaneously
in all these places.

13. The best method of catching the animals is by trapping,
but this method must be improved. It seems probable that live
bait is not attractive to the animals. Instead of relying upon bait,
the large openings of the traps should be laid in the path of the
animals.

I desire to express my indebtedness and warm thanks for the
assistance which they have so generously rendered me in this
work, to the persons whose names I take pleasure in stating
below :

Mr. C. A. Campbell, instructor in biology at Coronal Institute
in San Marcos, for his enthusiastic and skillful assistance and
his most enjoyable company during the collecting trips.

Dr. H. F. Moore, Acting Commissioner of U. S. Fish Hatch-

OBSERVATIONS ON TYPHLOMOLGE RATHBUNI. 103

cries, for permission to keep apparatus and outfit as well as ma-
terial collected at the U. S. Fish Hatchery in San Marcos.

Mr. Frank Johnson for permission to use and abuse his well in
every conceivable way and for valuable suggestions and help in
catching Typhlomolge out of the well.

Mr. Mark Riley, Superintendent of U. S. Fish Hatchery in
San Marcos for aiding the work in every possible way.

Dr. W. T. Vaughan, of the U. S. Geol. Survey, and Prof. C.
Eigenmann for valuable suggestions as to traveling and local con-
ditions in Texas.

Dr. T. W. Stanton and Mr. L. W. Stephenson, of the U. S.
Geol. Survey, for determination of the various rock specimens
collected from the caves.

Dr. H. E. Ortmann for identification of the various species of
Crustaceans.

Dr. L. Stejneger for identification of the salamander Plethodon
glntinosiLS.

Library of the Brooklyn Museum and in particular Miss S. H.
Hutchinson, the librarian, for extensive help in obtaining the
literature.

Mr. S. W. Stanfield, teacher in biology at the State Normal
School in San Marcos for valuable suggestions.

BIBLIOGRAPHY.

Benedict, J. E.

'16 Preliminary Descriptions of a New Genus and Three New Species of
Crustaceans from an Artesian Well at San Marcos, Texas. Proc. U. S.
Nat. Mus., xviii, 615.
Eigenmann, C. H.

'09 Cave Vertebrates of America : A Study in Degenerative Evolution.

Publication No. 104 of Carnegie Institution, Washington.
Emmerson, E. T.

'05 General Anatomy of Typhlomolge rathbuni. Proc. Soc. Nat. Hist., Bos-
ton, xxxii, 43.
Hill, R. T., and Vaughan, T. W.

'13 Geology of the Edwards Plateau and Rio Grande plain adjacent to
Austin and San Antonio, Texas. With reference to the occurrence of
underground water. i8th Annual Report, U. S. Geol. Survey, Part II.,

193-
Normann, W. W.

'oo Remarks on the San Marcos Salamander, Typhlomolge rathbuni, Stejne-
ger. Amer. Nat., xxxiv., 179.

104 EDUARD UHLENHUTH.

Stejneger, L.

'96 Description of a New Genus and Species of Blind Tailed Batrachians
from the Subterranean Waters of Texas. Proc. U. S. Nat. Mus., xviii,
169.
Weckel, A. L.

'07 The Fresh-water Amphipods of North America. Proc. U. S. Nat. Mus.,
xxxii, 25.

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 7

ON THE DEVELOPMENT OF THE SPONTANEOUSLY

PARTHENOGENETIC EGGS OF ASTERINA

(PATIRIA) MINI ATA.

H. H. NEWMAX.i

INTRODUCTION.

While engaged in a series of experiments on echinoderm hy-
bridology, which I was conducting during the months of April
and May, 1920, at Pacific Grove, California, I was forcibly struck
by the frequency with which spontaneous parthenogenesis occurs
in the starfish, Astcrina (Patina*) minata. When I use the term
" spontaneous," I mean that eggs were in no way treated either
by physical or by chemical agents. Precautions were taken,
moreover, to prevent accidental fertilization. The procedure was
as follows :

Sea water, brought in from the open sea and therefore free
from the chemical impurities present in sea water that has been
pumped through metal pipes, was allowed to stand at least four
days at laboratory temperatures, which during the month of May
ranged from 16° to 19° C. It is certain that no sperms could
live for this length of time in sea water. This method is chosen
in preference to Loeb's practise of heating the sea water to 60°
C. for some time, because it involves no possible chemical
changes in the sea water nor any driving out of oxygen. Before
opening a starfish, it was scrubbed thoroughly in cold running
fresh water and rinsed in a strong stream of fresh water. In
case the animal proved to be a male it was discarded, and hands
and instruments were scrubbed in fresh water before touching
another starfish. If the animal proved to be a ripe female the
ovaries were gently shaken into a finger-bowl containing 150 c.c.
of the sea water prepared for the purpose and the bowl was cov-
ered with a clean glass plate and placed upon a table out of reach

i From the Hopkins Marine Station of Leland Stanford, Jr. University
and the Hull Zoological Laboratory of the University of Chicago.

105

IO6 H. H. NEWMAN.

of direct sunlight. For purposes of observation eggs were re-
moved from time to time and placed in watch-glasses with a ster-
ilized pipette. It was thought safer not to run, as direct controls,
normally fertilized eggs, but on other days and under identical
conditions numerous observations were made on the course of
normal fertilization and development. The differences observed
between the behavior of parthenogenetic and that of fertilized
eggs were so striking that it seemed well worth while to study
and to describe them.

PECULIARITIES OF THE MATERIAL USED.

Asterina (Patiria) miniata is one of the commonest starfishes
of the California coast. It is a relatively small species in which
the five rays are almost completely amalgamated with the central
disc in such a way as to give the creature a nearly pentagonal out-
line. In color it ranges from a brilliant scarlet to a light cream
color with various intergrades and piebald combinations. It
seems likely that there are several subspecies that freely inter-
breed. My experience seems to indicate that relatively few eggs
mature at a time and are exuded in small numbers over a long
season. Among the very large number of females examined I
never found an ovary that showed any large percentage of ripe
eggs. Those that were shed from the removed ovary and gently
shaken in sea water contained oocytes in all stages of develop-
ment, some quite small, others fully grown but incapable of mat-
uration, still others mature and ready for maturation and fer-
tilization after standing about one and a half to two hours in sea
water. None of the eggs when freshly shed had undergone matu-
ration. This may mean that the artificial shedding of the eggs
is a premature process, and that, as seems to be the case in some
asteroids, if the eggs were to be normally extruded through the
genital pores, they would be immediately ready for fertilization.
I have never been able to observe the extrusion of eggs in As-
terina, nor to induce it by massaging, as can be done in some
asteroids. This may possibly be due to some peculiar sexual
rhythm in this species, which results in ovulation occurring only
at night or at some particular phase of the moon or of the tide.

EGGS OF ASTERINA (PATIRIA) MINIATA.

On this point I have no evidence. So it should be borne in mind
that the eggs used in these experiments, and presumably by Loeb,
who did a considerable amount of work upon artificial partheno-
genesis in this species, are not the normally shed eggs equivalent
to those which are fertilized in nature, but are probably prema-
turely shed eggs. It is very questionable, therefore, whether
parthenogenesis ever occurs in nature. It seems more probable
that the parthenogenetic eggs observed in these experiments re-
sult from the artificial conditions involved in shedding some of
the eggs prematurely.

The eggs of Asterina are very hardy and resistant of the cyto-
lytic action of sea water. While the unmaturated oocytes of most
echinoderms begin to disintegrate within twenty-four hours,
those of Asterina frequently remain unchanged, as though in
stable equilibrium, for from four to eight days. I have before me
a number of microscopic whole mounts, showing numbers of
eggs and larvae of Asterina fixed in Bouin's solution on the eighth
day after fertilization, in which there occur a number of un-
matured oocytes with germinal vesicle clean-cut and spherical
and plasmosome well defined. This ability of the unripe eggs to
withstand disintegration is of great practical value in the study
of development, for the water is kept free from the products of
egg decay ; a decided advantage in view of the fact that where
there are small percentages of developing eggs surrounded by
large percentages of non-developing eggs, the former would have
very small chance of survival if the latter were to decay and foul
the water.

EXPERIMENTAL DATA.

Most of the detailed data on pathenogenesis in Asterina were
obtained during the month of May, 1920, although the phenom-
enon was noted incidentally throughout April. The month of
May seems to be the best month for work with Asterina as there
appear to be larger ovaries and more full-grown oocytes then
than earlier. Possibly June would be still better, though I have
not tried any experiments at that time.

Some thirty-two experiments were made in all, and no two
gave exactly identical results. A large series of ten experiments

108 H. H. NEWMAN.

made on May 20 shows the full range of diversity and will serve
to illustrate all of the points of interest. As the time element is
of prime importance in this study it is necessary to give readers
as a norm for purposes of comparison a time schedule of the
development of the normally fertilized Asterina.

TIME SCHEDULE OF THE DEVELOPMENT OF NORMALLY FERTILIZED EGGS OF

ASTERINA.

Hours after Shedding. Condition of Maturated and Fertilized Eggs and Embryos.

2*4 Fertilization membranes formed on majority of matured eggs.

3*^ Cleavage beginning.

5 Cleavage taking place in all fertilized eggs, a few 4 and 8

cell stages.

7 Many stages as advanced at 32 -f- cells.

9 Most of eggs in early blastula stages, but there are a good

many eggs without membranes, in early cleavage stages.
These are probably parthenogenetic.

25 Two distinct types of larvae present : the great majority being

typical gastrulae, swimming up near the surface of the dish,
and capable of being readily pipetted off; a small minority
of living larvae variously abnormal and swimming at or
near the bottom. These latter are probably, though not
certainly, parthenogenetic.

49 Normal larvae, early bipennariae, forming enterocoel pouches.

73 Bipennariae with well-defined ciliated bands.

96 Advanced bipennarise with mouth, oesophagus, stomach, in-
testine, anterior and posterior enterocoels, waterpore, etc.

TIME SCHEDULE OF THE DEVELOPMENT OF PARTHENOGENETIC EGGS OF ASTERINA.

No. of Ex- Hours after
periment. Shedding. Condition of Maturated Eggs or of Developing Embryos.

1 7 No membranes and no cleavage.

9 About i per cent, of eggs show distinct membranes,

no cleavage.

24 About y2 of i per cent, swimming blastulae, all sub-
normal in appearance, irregular in shape or solid.
About 2 per cent, early cleavage stages (2 and 4
cell stages).
48 All larvae and cleavage stages undergoing cytolysis.

II 7 No membranes and no cleavage.

9 No membranes and no cleavage.

24 No larvae nor cleavage.

Ill €>y2 No membranes and no cleavage.

8% No membranes, but about il/2 per cent, of cleavage

stages, mostly fairly regular 2 and 4 cell stages.
25 Nearly i per cent, blastulae, mostly solid and mo-
tionless.

EGGS OF ASTERINA (PATIRIA) MINIATA. 109

IV 6 No membranes and no cleavage.

8% About 2 per cent, of eggs cleaving without mem-
brane formation, some as advanced as 8 cells.

25 Nearly 2 'per cent, blastulae, slightly abnormal and

motionless.

V 6l/2 About 5 per cent, of eggs with rather narrow but

distinct membranes, no cleavage.

•jT/2. About il/2 per cent, of eggs without membranes in

early cleavage stages (2, 3, 4 cells), some quite
regular, but the majority more or less irregular.

8J^ About 5 per cent, of eggs cleaving, ranging from 2

to 1 6 cells. Evidently a good deal of cleavage
has begun since the 7^2 hour observation. A few
of the eggs now show wide typical membranes
and there are transitional stages between these
and the more plentiful type with narrow mem-
branes.

26 About 7 per cent, swimming blastulse, some nearly

normal, but the majority solid, wrinkled or other-
wise abnormal. Eggs with membranes, under-
going black cytolysis.

31 Many gastrulaa, some with two or more archentera

some exogastrulae, etc.

74 A good collection of twin larvae, studied in a subse-
quent connection. No normal larvae. All larvae,
swimming on the bottom of the dish.

VI 6% About 4 per cent, of eggs with distinct but narrow-
membranes, no cleavage.

8% About 2^/2 per cent, of eggs with wide, typical

membranes, but no egg with membrane shows
cleavage; about ^2 of i per cent, of eggs showing
cleavage stages, ranging from 2 to 8 cells, some
quite regular.

2&l/2 A very few larvae, all subnormal, some motionless,

others feebly swimming.

VII 7 No membranes, no cleavage.

9 No membranes, but about 2 per cent, of eggs in

cleavage stages ranging from 2 to 32 cells.

z6l/2 About i per cent, subnormal blastulae, a few swim-
ming.

32 A very few larvae undergoing gastrulation ; several-

with two or more archentera.

VIII 6^2 No membranes, no cleavage.

9 No membranes, but about 3^2 per cent, of eggs in

cleavage stages, ranging from 2 to 16 cells.
26 About 2l/2 per cent, of larvae, all subnormal.

HO H. H. NEWMAN.

IX 7 About 8 per cent, of maturated eggs with narrow

membranes ; about 3 per cent, of eggs without
membranes showing first steps in cleavage, a few
having completed the first cleavage.

Sy2 A few eggs in 4- and 8-cell stages.

3ol/2 A fraction of i per cent, of larvae undergoing gas-

trulation, and swimming about, the best of them
being nearly normal in appearance.
78 All larvae dead.

X 7*4 About 5 per cent, of eggs with fully typical mem-
branes, no cleavage stages.

§1/2 Over 50 per cent, of maturated eggs in cleavage

stages, without membranes, ranging from 2 to 8
cells. No eggs with membranes segmenting.

31 About 75 per cent, of all maturated eggs have un-
dergone cleavage without membrane formation,
and are in various stages ranging from early
cleavage to gastrulse. Numerous dwarf blastulae,
due to blastolomy ; many gastrulse with plural
archentera ; solid blastulae and exogastrulas in
considerable numbers. All were swimming about
on the bottom of the bowl. This culture was
made the basis of a study of the more advanced
development of parthenogenetic eggs and will be
referred to in more detail in a subsequent dis-
cussion.

SIGNIFICANT POINTS BROUGHT OUT BY THE DATA SHOWN IN

THE ABOVE SCHEDULE.

1. Membrane formation occurs in exactly half of the experi-
ments here described. It was observed in from i to 8 per cent,
of the maturated eggs.

2. The degree of completeness of membrane formation varies
greatly in different sets of eggs and in different eggs of a given
set. In some eggs the membrane is so little lifted from the sur-
face of the egg as to be scarcely noticeable, ibut in others the
membrane is indistinguishable from that seen in fertilized eggs.

3. Eggs that form membranes, whether narrow or wide, do not
further develop, but undergo cytolysis within twenty-four hours.

4. The percentage of maturated eggs that undergo partheno-
genetic development varies from none to about seventy-five; in
experiment II no cleavage occurred, while in experiment X 75
per cent, of all maturated eggs at least began cleavage. The aver-

EGGS OF ASTERINA (PATIRIA) MINIATA.

Ill

age number of parthenogenetic eggs is about two per cent, of
maturated eggs.

5. Eggs that undergo parthenogenetic cleavage never form
" fertilization " membranes, the closely fitting vitelline membrane
being the only envelope that surrounds the blastomeres.

6. Cleavage in parthenogenetic eggs never begins earlier than
six and one half hours after the eggs are placed in sea water and
the average time for the beginning of cleavage is about seven and
a quarter hours. Cleavage begins in fertilized eggs sometimes
as early as three and one half hours, and the average time of be-
ginning is about four hours. Subtracting two hours for matura-
tion to complete itself, we have cleavage beginning five hours
after maturation in parthenogenetic eggs and two hours after
maturation in fertilized eggs. There is, therefore, a retardation
in development in the case of parthenogenetic eggs of three hours,
and at a very critical period.

2 4 6

FIGS. 1-6. Cleavage stages in spontaneously parthenogenetic eggs of As-
terina. i. An incompletely segmented two-cell stage in which one blastomere
is in advance of the other. 2. A normal two-cell stage. 3 and 4. Irregular
cleavage stages. 5. A typical normal four-cell stage. 6. A rather common
type of abnormal cleavage in which one blastomere is undergoing cytolysis
and the other is remaining normal.

7. Cleavage and subsequent development in parthenogenetic
eggs take place much more slowly than in fertilized eggs. Even
the most nearly normal parthenogenetic eggs take nearly twice
as long to reach a given stage as do fertilized eggs. Develop-
ment is, therefore, greatly retarded and we would naturally ex-

112 H. H. NEWMAN.

pect the larvae to exhibit the various types of developmental de-
fects that are commonly seen in inhibited individuals.

8. In none of the numerous experiments did parthenogenetic
'eggs give rise to even approximately normal bipennariae. The
most successful larvae were certain double monsters that will be
•discussed later.

9. The average viability of parthenogenetic larvae varies greatly
in different sets. As a rule viability was lowest in those sets in
which the smallest percentage of larvae occurred and highest in
those in which the largest percentage of larvae occurred.

10. Individual viability varies greatly within a given set of
eggs. Quite frequently eggs die and disintegrate during the first
or subsequent cleavages, while it was not uncommon for a few
larvae in each of the best sets to live for from four to seven days.

11. Cleavage in parthenogenetic eggs is sometimes very normal
in appearance, but in every set the majority of cleavage stages
are irregular (Figs. 1-6). Sometimes blastomeres of the two
cell stage separate, and form half-sized blastulae, seldom going
further. In other cases one or more blastomeres cease cleavage
while the rest go on and form a covering of small cells about a
large central cell. Numerous other cleavage anomalies occur
which need not be detailed here.

DISCUSSION.
Loeb's Observations of Spontaneous Parthenogenesis in Asterina.

Doubtless the reader recalls the work of Loeb (1905) on
"Artificial membrane formation and chemical fertilization in a
starfish (Asterina)." In this paper the author describes various
methods employed first, for inducing membrane formation and
second, for inducing cleavage and subsequent development in the
same species of starfish which forms the material of the present
investigation. Loeb recognizes the occurrence of spontaneous
parthenogenesis in Asterma as is shown by the following quota-
tions : " The eggs of the starfish show a slight tendency to develop
spontaneously without any external influence." " If the eggs of
Asterma are allowed to mature in sea water and are left to them-
selves, sometimes none, sometimes a fraction of a per cent., some~

EGGS OF ASTERINA (PATIRIA) MINIATA. 113

times more, will segment and develop into larvae. But the de-
velopment of these eggs is much slower than that of fertilized
eggs and, as a rule the larvae are not so perfect and die sooner."
"We have, therefore, two types of development in these (As-

i

terina) eggs. One type is represented by the fertilized egg, and
this type can be producecf artificially in a number of eggs, at least,
by calling forth the membrane formation by the above-named
artificial means. The second type is represented by the spon-
taneously developing egg in which no membrane has been called
forth; these latter eggs begin to segment later, and possibly
develop more slowly than the other eggs, and form larvae which
are not as perfect as those belonging to the first type."

It will be seen that Loeb has touched upon some of the essential
points that are brought out in my experiments. He notes that
spontaneous parthenogenesis occurs in a small per cent, of eggs ;
that parthenogenetic cleavage takes place without membrane
formation ; that cleavage begins later ; and that development is
slower and less normal than is fertilized eggs. Loeb, however,
was not primarily interested in the course or results of spon-
taneous parthenogenesis, but merely dealt with it incidentally as a
check upon his work on artificial parthenogenesis or chemical
fertilization. He, therefore, merely points out the foregoing par-
ticulars without entering into any discussion as to their
significance.

Spontaneous Membrane Formation.

In only one important point is there lack of essential agreement
between his results and mine : he failed to note any cases of spon-
taneous membrane formation which was so frequently noted in
my experiments. I am at a loss to explain this discrepancy
between his results and mine, both performed at Pacific Grove
and both unquestionably safeguarded against accidental error.
Possibly the material behaves differently at different times of the
year and Loeb's work was done at quite a different time from
mine, which was confined to the last few days of April and the
first three weeks of May. The only difference in treatment be-
tween Loeb's cultures and mine had to do with the methods of

114 H> Hl NEWMAN.

sterilizing the sea water in order to avoid normal fertilization.
Although he does not mention the fact in these particular experi-
ments, his practice was to use boiled or highly heated sea water ;
while I used sea water that had been kept in a demijohn for at
least four days. It seems barely possible that heating of the
water prevents spontaneous membrane formation. I would, of
course, have tried this experiment, had I known of Loeb's detailed
paper at the time of my experiments, but I had with me only his
book on " Artificial Parthenogenesis and Fertilization " and in
that book he fails to mention the occurrence of spontaneous
parthenogenesis in Asterina. Heated sea water would doubtless
be relatively poor in oxygen and this might be responsible for his
failure to find spontaneous membrane formation. It is possible
also that this process (membrane formation) was so belated in
its appearance that it occurred after Loeb had ceased to look for
it in his cultures. Unless one gets a very early start in experi-
ments with this material the working day is likely to be over
before any signs of membrane formation appear, and the next
morning these eggs will have undergone cytolysis and their mem-
branes will have disappeared. My plan was to make a before-
breakfast expedition to the collecting grounds, get the material
ready for work, breakfast, and be ready for experimentation by
8:00 A.M. If one begins to collect after breakfast it is likely to
be nearly 11:00 A.M. before experiments with Asterina eggs
could be commenced. If such were the case it would be 6:00
P.M. before distinct membranes would be visible, later than the
time when the investigator habitually " knocks off for the day."
It is barely possible then that Loeb may have missed spontaneous
membrane formation in some such way as this.

Of the validity of my obervations there seems to be no doubt.
In answer to the criticism that failure to heat the sea-water
vitiates the observations, it may be said that if these eggs with
membranes are fertilized, they should segment ; but they never do.
We seem to be forced to the conclusion, therefore, that membrane
formation in Asterina, which Loeb has been at such pains to bring
about by chemical means, occurs spontaneously in a considerable
percentage of cases. This being true, the various manipulations

EGGS OF ASTERINA (PATIRIA) MINIATA. 115

used by Loeb have merely served to hasten a natural process and
to cause it to occur in a larger percentage of eggs.

Something in addition to membrane formation occurs in the
eggs handled by Loeb, for they go ahead and segment as do
normally fertilized eggs, while the eggs that form spontaneous
membranes do not segment. From this it would appear that the
so-called "fertilisation membrane" is not an essential feature of
development, but merely its usual accompaniment. Thus the
voluminous literature dealing with artificial membrane formation,
as though it were the most important event in the initiation of
development, loses some of its force. Loeb himself recognized
that development, at least in Asterina, could proceed without
membrane formation ; witness his statement, corroborated by my
own observations, that spontaneous parthenogenesis proceeds
without the preliminary of membrane formation. He seems to
suggest, however, that this is not to be considered as typical de-
velopment, since it begins later, goes more slowly and results less
normally than in the case of fertilized eggs. Exactly similar
results may be obtained, however, in fertilized eggs by the use of
agents that retard development, such as cold, hybridization,
anaesthetics, etc. So we must admit that real development may
occur without membrane formation, and that membrane forma-
tion may occur without initiation of development. The two
processes are independent though they usually are associated
in normal ontogeny.

The Development Spontaneously of Partheno genetic
Eggs of Asterina.

According to Loeb, the development of the chemically fertil-
ized eggs differs from that of the spontaneously fertilized eggs
in two respects ; first, in forming membranes ; and second, in be-
ginning earlier and proceeding more rapidly. With this distinc-
tion I fully agree. Evidently, in the chemically fertilized eggs,
something in addition to membrane formation takes place, a
something that results in prompt initiation of the changes ex-
pressed by cleavage. In the spontaneously parthenogenetic eggs,
however, initiation to development is very slow in beginning, and

Il6 H. H. NEWMAN.

is less effective when it does begin. In last analysis the difference
is evidently to be expressed in terms of rate of change. The
whole process of ontogeny in these eggs is from the beginning
retarded, and the results are exactly similar to those which may
be obtained by the use of growth-retarding agents applied to newly
fertilised eggs. The first effect of pronounced retardation of the
normal growth process in the egg is the partial or complete
obliteration of the characteristic axes of polarity and symmetry
in the egg, a breaking down of the axial gradient. This results
subsequently in loss of unity of organization, involving physio-
logical isolation of blastomeres or of cell aggregates, in double
and triple polarity and consequent double or triple monsters, and
in a whole series of products of differential inhibition, such as
those described by Child for sea-urchin.

8 10 1Z

FIGS. 7—12. Later developmental stages in spontaneously parthenogenetic
eggs of Asterina. 7. The commonly occurring solid blastula type. 8. A pair
of twin blastube enclosed within one vitelline membrane, evidently the result
of physiological isolation of two blastomeres in the two-cell stage. o. A mul-
tipolar embryo gastrulating at several points. 10. Double monster with two
symmetrical archentera. //. Another double monster with an additional an-
terior archenteron. 12. A microcephalic ciliated larva, with differentiated
stomach and intestine, but no anterior parts.

Some of the types of inhibited larvae found in cultures of spon-
taneously parthenogenetic Asterina eggs are shown and described
in figures 7-12. The solid blastula is the commonest type, a type
devoid of an axis of polarity (Fig. 7). Forms that are bipolar

EGGS OF ASTERINA (PATIRIA) MINIATA. 1 17

and tripolar, etc., usually undergo gastrulation in two or more
places (Figs. 9, 10, n) and produce double and triple monsters,
etc. The most nearly normal forms are decidedly abnormal
early bipennaria larvae in which the anterior parts are relatively
inhibited. Such a form is shown in Fig. 12, in which the mouth
never breaks through, oesophogus is not clearly differentiated,
but stomach and intestine are well devevloped. Since I have in
preparation a detailed paper on twinning in Asterina, in which I
intend to discuss the physiology of twinning in general, I shall not
enter further into an account of ' the various types of inhibited
larvae that result from spontaneous parthenogenesis, for the same
types result from several other kinds of inhibiting factors. In
closing I merely wish to emphasize this one point : that the results
of spontaneous parthenogenesis are those usually found to accom-
pany early grozvth retardation. For a summary of this paper the
reader is referred to the eleven points referred to on pages 110-
112 and to the italicized clauses in the general discussion.

I am greatly indebted to the Hopkins Marine Station of Le-
land Stanford University, and to its director, Dr. Walter K.
Fisher, for the excellent facilities for research that I enjoyed
while at Pacific Grove.

LITERATURE.
Loeb, J.

'05 Artificial Membrane Formation and Chemical Fertilization in a Star-
fish (Asterina) Univ. of Calif. Pub., Physiol., Vol. 2, No. 16.
'13 Artificial Parthenogenesis and Fertilization. Univ. of Chicago Press.

AUTHOR S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 7

ON THE OCCURRENCE OF PAIRED MADREPORIC

PORES AND PORE-CANALS IN THE ADVANCED

BIPENNARIA LARVAE OF ASTERINA (PATIRIA)

MINIATA TOGETHER WITH A DISCUS-

SION OF THE SIGNIFICANCE OF SIMI-

LAR STRUCTURES IN OTHER

ECHINODERM

H. H. NEWMAX.i

INTRODUCTION.

Ever since the theory became current that the bilaterally sym-
metrical larvse of echinoderms afford phylogenetic evidence that
this group of radially symmetrical animals was derived from
bilaterally symmetrical ancestors, larvse that showed more than
the normal tendencies toward persistent bilaterality have had a
special significance.

Normally, the first evidence of the encroachment of the adult
radial symmetry upon the the larval bilaterality is seen in the
development of a distinct hydroccel with a madreporic pore and
pore-canal on the left side, and none on the right. This failure
of the right side to keep pace with the left has been considered
as the mechanical cause for the twisting around of the serially
repeated primordia of the radial water canals and the assumption
of the adult radial symmetry.

The occurrence, therefore, in the larvse of at least two classes
of echinoderms, of paired right and left madreporic pores, pore-
canals, and other derivatives of the hydrocoels, looks like the per-
sistence of the ancestral bilateriality and tends to strengthen the
current theory as to echinoderm phylogeny.

As early as 1892 Field, in his account of the larval development
of Asterias vulgaris, described the transitory appearance in all

i From the Hopkins Marine Station of Leland Stanford University and
the Hull Zoological Laboratory of the University of Chicago.

118

BIPENNARIA LARVAE OF ASTERINA (PATIRIA) MINIATA. 1 19

larvae aged about three and a half days of a strictly bilaterally
symmetrical condition of the hydroccels, madreporic pores, and
pore-canals. Soon after this period the right madreporic pore
closes and the pore-canal, without an external opening, persists
for a short time and then entirely disappears.

Several continental writers had observed the occasional occur-
rence in asteroid larvae of paired madreporic pores, but had con-
sidered the condition as a pathological one. Field, however,
maintains that there normally occurs, in Asterias vulgaris at least,
a transitory stage in which bilateral madreporic pores exist, and
considers this condition as " a true ontogenetic character, and of
very considerable phylogentic significance." Gemmill (1912) was
able partially to confirm Field's observations of the occurrence of
paired water pores in the genus Asterias, finding this condition in
fifty per cent, of larvae of Asterias glacialis and in about ten per
cent, of those Asterias rubcns. He also states that in all cases
the water-pore soon closes.

Paired echinus-organs, madreporic vesicles, and other deriva-
tives of the right hydroccel, were observed by MacBride (1911)
in two very advanced Plutei, one of Echinus millaris and the other
of Echinus esculentis. A similar condition, was observed by
Grave (1911) in a single Pluteus of the sand-dollar Mcllita penta-
pora, though the figure shows the two madreporic canals making
exit through a single median dorsal madreporic pore.

PAIRED MADREPORIC PORES IN ASTERINA LARVAE.

During the months of April, May and June, 1920, I had occa-
sion to observe the development of a very large number of cul-
tures of larvae of Asterina (Patina) miniata at Pacific Grove,
California. In one culture consisting of otherwise normal,
healthy larvae three weeks old, I noticed a larva with two perfect
madreporic pores and pore-canals. Further search, involving a
complete census of all the larvae in the culture, revealed twenty-
six more larvae with right madreporic canals in some cases as well
developed as the left ; but in others with the right pore closed and
the pore-canal smaller than the one on the left. More than half
of all the advanced larvae in this culture had the double madre-

I2O

H. H. NEWMAN.

poric pores or at least double pore-canals, while the remainder of
the larvae were quite typical. After a prolonged search in other
cultures of Asierlna, I was unable to discover any other larvae
with double madreporic pores.

These double-pored larvae were .carefully watched and attempts
were made to rear them by introducing diatoms collected from

FIG. i. Dorsal view of largest bipennaria, viewed as a transparent object.
The hydroenterocoel cavities are shown stippled. Note both left and right
madreporic pores, opening far apart ; distinct pore-canals on both sides.

the kelp found in the tide pools inhabited by Asterina. They fed
to some extent on several species of diatoms, but made no progress
beyond the condition in which they were when the double-pored
character was first noticed. Instead, they slowly retrogressed in

BIPENNARIA LAEVJE OF ASTERINA (PATIRIA) MINIATA. 121

differentiation, became sluggish, and died without having begun
to metamorphose. Throughout their lives they remained bilater-
ally symmetrical and observations made at the end of the fourth
week, shortly before the culture began to die out, showed that the
right madreporic pore was still open in several of the largest bi-
pennarise. While the larvse were still active and in good health,
several specimens were quieted with chloretone and mounted for

2 3

FIG. 2. Dorsal view of a somewhat less advanced bipennaria with madre-
poric pores close together but both open.

FIG, 3. Dorsal view of one of the least advanced bipennariae with right
madreporic pore closed, but with distinct pore canal on the right side.

microscopic study. These were drawn with camera lucida and
three of these drawings are shown in Figs. 1-3. The largest
larva in the culture is shown in Fig. I. Its length was a little
over 7 mm. The anterior hydroenteroccel pouches have become
completely fused in front of the oral funnel and have grown
forward into a median pre-oral ccelomic cavity. The right and
left body cavities are equally well developed throughout and each
has grown posteriorly beyond its own side and has curled round

122 H. H. NEWMAN.

the intestine, so as to have its blind end pointed forward on the
opposite side of the body. Thus these two posterior outgrowths
cross one another and encroach on each other's territory, a condi-
tion that could not well represent any ancestral state, but one to
be expected in a double monster such as I believe this to be. Each
madreporic canal connects broadly with the dorsal wall of its
respective hydroccel, slopes slightly toward the median line, and
opens by a distinct pore to the exterior. The two pores are not
very close together in this specimen. The other two larvae are a
little smaller and less advanced than that shown in Fig. i. That
shown in Fig. 2 is about 6 mm. in length, has no pronounced for-
ward growth of the preoral ccelom and does not show an over-
lapping of the posterior extensions of the right and left cceloms.
The pores of the two pore-canals are very close together and
might form one double madreporite. The specimen shown in
Fig. 3 was of about the same stage of advancement as that in
Fig. 2 but differs primarily in the fact that the right madreporic
pore is closed and its canal is smaller and shorter than its left
partner. Also the anterior cceloms have not fused.

In discussing these anomalous larvae with Dr. Walter K. Fisher,
director of the Hopkins Marine Station, I was interested to learn
from him that he had, while collecting, noted adult specimens of
Asterina, and of other species of asteroids, in which there were
two madreporic plates and, correspondingly, two stone canals.
This information immediately suggested to me the strong prob-
ability that these adults with paired madreporic plates must
have arisen from larvae with double madreporic pores such as I
had under observation. The question would then arise as to
whether this double condition in the adults would bear the same
phylogenetic interpretation as has been offered for the double
condition in the larvae. If not; why not? But this involves us
in a discussion of the significance of these anomalous paired struc-
tures in echinoderms.

DISCUSSION.

Three distinct interpretations have been offered for the appear-
ance of these anomalous right-hand elements that normally
appear only on the left.

BIPENNARIA LARVJE OF ASTERINA (PATIRIA) MINIATA. 123

1. The first interpretation is well phrased by MacBride1 as
follows :

" The appearance of a right and left madreporic pore is the first
Indication of zvhat is really the key to the understanding of Echi-
noderm development, vis., the fact that the tivo sides of the larva:
originally gave rise to precisely similar organs, but that sonic of
these organs grew and developed on the left side while they
atrophied on the right, and that thus an asymmetry was
produced."

This seems to be the natural interpretation of the facts, but
there are certain other facts that this interpretation does not
cover. In a larva of the sand-dollar Mcllita pcntapora, Grave
(1919) found that not only were the mesodermal structures, in-
cluding hydrocoel, pore-canals, etc., bilaterally repeated, but also
paired ectodermal pouches occurred which had no reference to the
water-vascular system, but were the primordia of the nervous
system. Grave is inclined to doubt the validity of interpreting
this extra ectodermal pouch as a reversion to an ancestral condi-
tion, and I would fully agree with him. The occurence of adult
starfishes with paired madreporic plates and stone canals also
weakens the phylogentic interpretation of double pores in larvae ;
for they are evidently strictly homologous structures. If the
double-plate condition in the adult is to be interpreted as a re-
version, of what is it an ancestral reminscence ? Surely not of an
ancestral starfish with radial symmetry of other organs, but bi-
lateral madreporic structures !

2. The second interpretation of paired madreporic structures
involves the idea that they are homceotic variations, in Bateson's
sense. Such variations may be viewed according to Grave " as
cases of perfected symmetry, the result of a long continued strain
due to imperfect balance, either morphological or physiological
or both, between the organism and its environment." I confess
that I am unable to see the force of this interpretation. It is
highly mystical in tone and savors of some internal perfecting
principle or " entelechy." If such a principle were operative, the
real problem would be to account for the failure of echinoderm

1 " Text-book of Embryology," Vol. I., p. 466.

124 H- H- NEWMAX.

larvae in general to continue to maintain their original bilaterality.
That is a different problem altogether.

3. A third theory involves the idea that these duplicated struc-
tures are the products of twinning and this is what I believe they
are. In laboratory cultures of Asterina (and in at least two
other species of asteroids with which I have worked) there are
numerous instances of twins and double monsters. The exact
cause of twinning and doubling is not fully known and will not be
discussed here, as I have in preparation a more extensive paper
on this subject. It may be said, however, that a long series oi
types has been found in which the original right and left pri-
mordia of the future embryo become more or less completely
physiologically isolated, and in proportion to the duration or com-
pleteness of the isolation, each half develops more or less inde-
pendently of the other. The result is a series of more or less
completely doubled larvae, as follows : twin blastulae which soon
separate and develop into half-sized larvae ; gastrulse with paired
archentera ; gastrulae with paired blastopores, but with archen-
tera fused anteriorly ;. early bipennariae with the anterior end of
the archenteron branched and with the beginning of a double set
of hyclroenterocoel pouches ; and finally advanced larvae with
double madreporic pores and water canals. I have placed the
type of anomalous larvae under discussion at the end of what
appears to me to be a logical series, representing the results of
varying degrees of physiological isolation of bilateral halves of
embryos. The type of result attained seems to depend on the
time of incidence of the cause or causes of the physiological
isolation in question and upon the degree of severity or duration
of the causal agent, whatever it may be.

There is therefore no more justification for the use of "double-
pored" larvae of echinoderms as evidence of an ancestral condi-
tion than there is for giving a similar significance to instances of
dicephaly or spina bifida in vertebrates. For they are all, in my
opinion, phases of "twinning" in the broad sense and will doubt-
less be explicable on some general physiological basis that we
shall hope to discuss in the future. Let me close with a word of
caution. Beware of giving a phylogenetic interpretation to an

r.ll'K. \.\.\RI A LARV-E OF ASTERINA (PATIRIA) MINIATA. 125

anomaly found in laboratory-bred larv?e of any sort. Such aber-
rations are more than likely to be the result of subnormal condi-
tions and nothing more.

LITERATURE.

Field, G. N.

'92 The Larva of Asterias vulgaris. Quart. Journ. Mic. Sc., Vol. 34.
Grave, C.

'n Metamerism of the Echinoid Pluteus. Johns Hopkins Univ. Circular,

New Series, 1911, No. 2.
MacBride, E. W.

'n Two Abnormal Plutei of Echinus in the Light which They Throw on
the Factors in Normal Development of Echinus. Quart. Journ. Mic.
Sc.. Vol. 57.
'14 Text-hook of Embryology, Vol. I. London.

Vol. XL. March, 1921. No.

BIOLOGICAL BULLETIN

DIFFERENCES IN VIABILITY IN DIFFERENT TYPES

OF REGENERATES FROM DISSOCIATED

SPONGES, WITH A NOTE ON THE

ENTRY OF SOMATIC CELLS

BY SPERMATOZOA.

JULIAN S. HUXLEY,
XEW COLLEGE, OXFORD.

Having occasion recently to go over the notes on which a pre-
vious paper of mine1 was based, I came across one result which,
although not published at the time, now seems worthy of record.

In the paper referred to, it was shown that by filtering chopped
Sycons through coarse instead of fine gauze, and pipetting off
the first-deposited portion of the cell-sediment, masses of cells
could be produced consisting entirely or almost entirely of cho-
anocytes.2 The unpublished data concern the occurrence, in cul-
tures of such collar-cell masses, of apparently normal regen-
erating masses similar to those obtained by straining through fine
gauze.3 This would appear to indicate that the chemotactic or
other attraction exerted by amcebocytes and dermal cells upon
each other is stronger than the similar attraction exerted upon
these same cells by choanocytes.4 This would lead to the ob-
served differential separation of the bulk of the dermal and
amoeboid cells in preparations where they were present only in
very small relative number.

These normal regenerates, as they may be called, in opposition
to collar-cell masses, were found in four of my twelve cultures

1 Huxley, Phil. Trans. Roy. Soc. (B), Vol. 202, 1911.

2 Ibid., p. 177.

3 Ibid., pp. 167-170.
* Ibid., p. 167.

127

128 JULIAN S. HUXLEY.

of collar-cell masses. In every case, they lived longer than the
collar-cell masses.

The time that elapsed before all the collar-cell masses (whether
spheres or blown-out masses) in any one of these four cultures
had shown the first sign of impending death by contracting, was
from 6 to 12 days; the time before all the masses of a culture
were dead, from 8 to 15 days. The time which elapsed before
the normal regenerates in the same culture died, however, was
from 14 to 33 days ; in most cases, all normal regenerates lived
longer than any collar-cell mass in the same culture. Whether
this greater viability of the masses containing all kinds of cells in
approximately normal proportions was due to a protective func-
tion exerted by the dermal cells after their migration to the ex-
terior, or to the fact of some dermal or amoeboid cells serving as
food for the rest, or to other possible causes, remains to be seen.
In any case, the facts are interesting.

A further observation may be referred to. It appears that the
spermatozoa of calcareous sponges have very rarely been ob-
served. Dr. Gatenby, of University College, London, who has
been working on the fertilization of sponges, was discussing the
subject with me, when I recalled that bodies resembling sper-
matozoa had been visible in some of my preparations of normal
regenerating masses from dissociations.

Some of my slides I lent to Dr. Gatenby, who re-stained them,
and was thus able to discover certain interesting facts. The facts
are briefly as follows : Round the margin of all cell-masses from
some of my experiments are to be seen minute deeply-staining
bodies resembling spermatozoa with an ordinary elongated head ;
and groups of such bodies are usually to be seen in the interior
of the preparations. They are, however, totally absent from
other slides representing other experiments. There can be no
doubt that these are spermatozoa. The interesting point about
them is that they swarm round the masses of cells whether these
contain oocytes or not. I.e., such sponge spermatozoa are at-
tracted by somatic cells as well as by their proper partners.

On re-staining, Dr. Gatenby found in the interior of many of
the somatic cells bodies which could be interpreted as heads of
spermatozoa which had been half converted into vesicular nuclei.

DIFFERENCES IN VIABILITY IN REGENERATES.

Here, however, it is impossible to give full proof of this until
further material can be examined. But it is at least suggestive
that Dr. Gatenby himself1 has found in the normal fertilization
of a closely-related sponge that the spermatozoa do not pene-
trate the oocytes directly, but enter collar-cells. Within these
they undergo a partial transformation to vesicular nuclei (at this
stage closely resembling the bodies found in the cells of the re-
generating masses), and are then transferred, by the migration
of the collar-cells, to the oocytes, into whose substance they pass.
Within the maturing ova they undergo the remainder of the trans-
formation to male pronuclei, and then effect fertilization in the
usual way. If the bodies within the cells of the regenerating
masses do prove to be what they appear to be, namely, half-
transformed sperm-heads, two interesting points emerge. The
first is that somatic cells can exert an attraction on spermatozoa
comparable to that exerted by oocytes. In normal fertilization,
only collar-cells within a certain radius are entered by sperma-
tozoa ; thus it might be supposed that the attraction was exerted
by the oocyte through the collar-cells, and that these had no at-
traction of their own. That this is not so, would be proved if
our interpretation of the bodies in the regenerating masses is
correct. But we would have to suppose that this attraction of
the collar-cells was much less than that of the oocytes, whose
presence thus would prevent collar-cells beyond a certain dis-
tance from oocytes from being entered.

•

In the second place, the definite but slight attraction of the
collar-cells for spermatozoa would be correlated with the definite
but partial transformation of the sperm-head to a nucleus within
them. This correlation between degree of attraction and degree
of nuclear transformation is what would be expected on such a
theory of fertilization as Lillie's.

It is to be hoped that any worker having the opportunity to ex-
amine ripe sponge spermatozoa will undertake an investigation of
the problem raised in this note.

In conclusion I have to thank Dr. Gatenby for his interest and
for permission to publish the results discovered by him.

1 " The Cytoplasmic Inclusions of the Germ-Cells. Part VIII. Fertilization
and Gametogenesis in Grantia compressa," Journ. Linnaean Soc., 1920.
May 28, 1920.

A MICRO-ELECTRODE AND UNICELLULAR

STIMULATION.

I. H. HYDE.i

During the summers of 1918 and 1919, while conducting ex-
periments on unicellular organisms and Echinoderm eggs, it was
found necessary to construct a micro-pipette that could be more
readily and more accurately controlled than could either Bar-
ber's1 pipette or Chambers's2 modification of the same.

An apparatus was needed with which it was possible, not only
to inject definite amounts of fluid, and extract special parts of
cells, but also with which these could be electrically stimulated.
Such an apparatus was devised, and although it was only in the
process of being perfected, nevertheless with it, fluid could be
injected and the membrane and other parts of Echinoderm eggs
extracted, and these as well as unicellular organisms electrically
stimulated. It was possible, for instance, to stimulate different
parts of Vorticella, and to determine that the contractile sub-
stance of the stalk of this special species differed from the con-
tractile substance of the frog's striated muscle fibers, in that, it
did not follow the law of " All or Nothing," which was discov-
ered for these muscle cells by Pratt4 with his very valuable and
interesting micro-electrode. The results pertaining to my ex-
periments with these devises will be published in the near future.

At present only the micro-electrode, imperfect though it be,
shall be described, in the hope that it will be improved by some
one who has the opportunity and is interested in the experimental
fields requiring its use.

The micro-electrodes are Barber pipettes, so modified as to be
employed for unipolar stimulation. They are glass tubes about
twelve centimeters long, six millimeters in diameter, and drawn

1 From the Woods Hole Marine Biological Laboratory and the University
of Kansas.

1 Barber, M. A., The Philippine Jr. Sc.. Sec. Trop. Med., 9, 307.

2 Chambers, R., Am. Jr. Pliys.. 1910, p. 189: BIOL. BULL., 1918, 34, p. 121.

3 Pratt. F.. Am. Jr. Pliys., 1917-1918-1919, Vol. 43. 44. 49.

130

A MICRO-ELECTRODE. 13!

out to a bent tip, having a lumen of from three or more microns
in the active electrode, and five or more times this in the indif-
ferent one. The other end of the glass pipettes are more or less
bent. If they contain mercury, they have a platinum wire sol-
dered into them, and are fitted airtight with rubber tubing and
clamps. The rubber tubing and clamp are supported on a pulley
that allows this end to be raised or lowered, thus aiding in the
adjustment of the mercury in the bent tip.

The pipettes are supported and regulated on the microscopic
stage of the Barber pipette holder, and their tips are operated
inside of a Barber moist chamber.

The pipettes may contain mercury or some electrolytic solu-
tion. Or the indifferent one mercury and the active one a solu-
tion, or partly mercury and only the tip a solution. Or both
pipettes may contain fine wire. The active one connected with
Pratt's glass-coated platinum tip sharpened to a point of 8 mi-
crons, and the indifferent one platinum. Or it may be platinum
and the active one contain an electrolyte. When non-polarizable,
Porter boot electrodes are introduced in the circuit, the pipettes
and the dish containing the boots may be filled with sea water.
Or by means of pressure applied from the rubber capped end of
the pipette, mercury can be forced toward the tip. Then by
diminishing the pressure it can be brought back into the capillary
a certain distance, drawing after it some of the solution desired
from a hanging drop in the moist chamber. An equilibrium is
then established that will remain constant as long as the condi-
tions are not changed. Or the active pipette electrode may be
filled with mercury either in accordance with Barber's method,
or filled under pressure as far as possible and then placed with
the indifferent electrode in contact with a mercury hanging drop.
If now the circuit from a battery is closed, the current enter-
ing the anode traverses the hanging drop and the cathode. At
the moment of the establishment of the current, the equilibrium
of forces that holds the mercury at a certain point in the capillary
is disturbed. The end of the fine thread of mercury moves up-
ward or downward a certain distance owing to a change in sur-

12

I. H. HYDE.

face tension. The direction and distance of the movement de-
pends of course upon the strength and direction of the current
introduced into the active pipette. When the mercury is brought
to the tip of this, the pressure clamp is closed, the mercury held
near the top, which may end in a drop of the same solution that
surrounds the organism that is to be stimulated.

If the active electrode be cathodal, a stimulus of a minimal
break shock will stimulate, for instance, the stalk or any desired
part of a Vorticclla, that is in the hanging drop, and near contact
with the active electrode, and the effect of the stimulation is
observed under the microscope.

A diagram of one micro-electrode is shown in Fig. I. The
pipette holder and moist-chamber are omitted in the illustration.

H

b H

d

FIG. i. a, battery; b, commutator; c, induction coil; d, clamp; e, platinum
wire; /, tip of pipette; g, clamp; /;, rubber tubing.

The movements of the mercury due to changes in surface ten-
sion by a force sufficient to overcome capillary attraction, in-
augurated by the passage of an electrical current, made it pos-
sible to employ this devise as a capillary electrode.

The fact that the meniscus of mercury moves more or less in
either direction in the tip of the micro-pipette by altering the
strength and direction of the current, and thus ejecting a solution
that lies between the mercury and the tip of the micro-pipette, or
drawing in a solution due to suction, led to the idea that it could
be adapted for the injection or extraction of minute quantities
of substances from unicellular structures. But the mechanism
needs further improvements before it can be satisfactorily em-
ployed for very accurate work.

I therefore was agreeably surprised while this paper was going

A MICRO-ELECTRODE. 133

to press, to find that Taylor1 devised an accurately controllable
micro-pipette, that seemed to fulfill the requirements of the
apparatus needed for injection and extraction. I believe that by
soldering plantinum wire into the capillary containing the mer-
cury and connecting it to batteries, comutator and coil, that it
will answer equally well for electrical stimulation.

1 Taylor, C. V., Science, 1920, June 18.

AUTHOR S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, FEBRUARY 28, IQ2I.

EFFECT OF VARIATIONS IN OXYGEN TENSION ON
THE TOXICITY OF SODIUM CHLORIDE ISOTONIC

TO SEA WATER.

ISAAC STARR, JR.,
MARIXE BIOLOGICAL LABORATORY.

Loeb first observed that he could protect fertilized Arbacia eggs
for some time from the toxic action of a NaCl solution, isotonic
to sea water, by removing the oxygen from the solution or by
adding a little KCN.1 Cyanide also protects unfertilized Arbacia
eggs against injury by isotonic solutions of various sodium salts,
and anaesthetics in appropriate concentrations have' a similar
effect.2 Experiments with Arenicola larvae have given similar
process being in some manner connected with the destructive
results.3 These results indicate the probability of an oxidative
process being in some manner connected with the destructive
reaction.

The following experiments were undertaken to determine the
effect of varying the oxygen tension on the toxicity of NaCl
isotonic to sea water. The form chosen to work with was the
larva of the marine worm, Arenicola cristata, a free swimming
trochophore of three body segments, about I/T, mm. in length.
\\ hen normal it is almost constantly in motion, showing ciliary
action and bendings to one side, or strong muscular contractions.
The larva? employed were all in the swarming stage and were
collected from the lighted side of the culture dish, just below the
surface of the water, where they gather in large numbers. As
they remain in this stage only from three to four days before
developing another segment and sinking to the bottom, a fairly
homogeneous culture is obtained. When these animals are placed
in pure sodium chloride isotonic to sea water they promptly con-

1 Loeb, J., Science, 1910, XXXII., 411; Biochemical Zeitschrift, 1910,
XXIX., p. So.

- Lillie R. S., Amer. Joitrn. Physio!., 1912, XXX., p. i.

3 Lillie, R. S., Amer. Joitrn. Physio!.. 1912, XXIX., p. 372, and 1913, XXXI.,
P- 255.

134

VARIATIONS IN OXYGEN TENSION.

135

tract to about -/$ of their former length and all motion ceases.
This contraction is followed by a slow relaxation and eventually
they are killed. But if replaced in sea water before death, they
slowly recover either completely or to a certain extent. The ratio
of those in motion to those still gives an index of the extent of
the injury.4

TECHNIQUE.

The larvae being positively heliotropic, gather in great numbers
in the lighted side of the culture dish. A pipette-full (over 500
larvae) was taken and placed in a watch glass, then, as the larvse
collected again on the light side, the water could be tilted off, the
last traces being removed by blotting paper. They were next
washed twice with the solution to be experimented with and then
transferred by a pipette to 50 c.c. of that solution. Then at
hourly intervals about 100 larvae were removed from the test
solution, washed twice with sea water and replaced in sea water.
These \vere inspected one hour and in most cases also twelve hours
after their return to sea water. Fifty individuals were examined
and the ratio of those in motion to those still was determined.

EXPERIMENTAL.

I. The larva? were first exposed to 0.52 molecular NaCl in
almost complete absence of oxygen. The solution was boiled
while a stream of hydrogen was passed through, care being taken

TABLE I.

EXPERIMENT i.
Over 50 larvae examined in each case.

Examined I Hour after Return to Sea Water.

Solution.

Time of Exposure.

3 Hours.

6 Hours.

Isotonic
sion . .

NaCl at atmospheric oxygen ten-

No motion

No motion

Isotonic

NaCl at reduced oxygen tension . .

All larvse show bend-
ings

No motion

4 For a further description of the behaviour of Arenicola larvse in isotonic
NaCl see Lillie, R. S., Am. Jour. Physio!., 1909, XXIV., p. 14, and Am. Jour.
Physiol.. 1911, XXVIII., p. 210.

136

ISAAC STARR, JR.

to maintain the proper concentration and reduce the temperature
to room temperature before introducing the larvae. This pro-
cedure removes the oxygen almost entirely, for if alkaline pyro-
gallol is added to the solution after this treatment no color ap-
pears for an hour or more. The control solution, 0.52 NaCl at
atmospheric oxygen tension, was also boiled, air instead of
hydrogen being bubbled through.

TABLE II.

EXPERIMENT i.
Over 50 larvse examined in each case.

A. Examined i Hour after Return to Sea Water.

Time of Exposure.

Solution.

i Hour.

2 Hours.

3 Hours.

Isotonic NaCl at at-
mospheric oxygen
tension

5 show feeble bend-

No motion

No motion

ings, ^ show no
motion

Isotonic NaCl at re-
duced oxygen ten-
sion

f show bendings

5 show bendings

No motion

more violent
than control

B. Same Larva: Examined 24 Hours after Return to Sea Water.

Isotonic NaCl at at-
mospheric oxygen
tension .

I larva showed All disintegrated All disintegrated
bendings. The
rest had disinte-
grated.

Isotonic NaCl at re-
duced oxygen ten-
sion . .

All larvae actively All disintegrated
motile

I All disintegrated

The antitoxic effect of the absence of oxygen is marked. This
was to be expected from Loeb's experiments on Arbacia eggs.

II. To expose the larvse to 0.52 mm. NaCl with increased
oxygen tension, the simple apparatus pictured was employed.
The oxygen tension desired was obtained by running washed
oxygen from a tank into the chamber " A," driving the level oT
water inside the chamber down to a mark previously determined

VARIATIONS IN OXYGEN TENSION.

137

by calculating the volume required with the necessary corrections.
The air already present in the chamber is thus diluted with pure
oxygen to the required degree. The water levels in and out of
the chamber were kept equal by removing water from the large
jar " F." The apparatus is only approximately accurate.

Bulb

Diagram of apparatus.

The larvae were introduced and removed by the tube "T" of
i mm. bore, which was connected to a Y tube and the rubber bulb
" H." By manipulating the rubber bulb and the two pinch cocks
on the connections of the Y tube, it was possible to pump the
larvae in or out of the small beaker " E." This beaker contained
the isotonic NaCl. A similar beaker "D" contained a control
culture in sea water at the same oxygen tension as " E." A third
beaker not shown in the diagram contained another control cul-
ture in isotonic NaCl at atmospheric oxygen tension.

In the first series of experiments the oxygen tension of the
NaCl solution containing the larvae was raised gradually. The
larvae were placed in the isotonic NaCl solution in the beaker
" E " before the oxygen tension was raised in the chamber. The
tension in the chamber was then raised to 320 mm. and the salt
solution and control were slowly stirred for two minutes. The
oxygen tension of the solution and control must rise slowly
towards 320 mm., but how long it is before it reaches that figure

138

ISAAC STARR, JR.

is difficult to predict. As it was impossible to stir too vigorously
for fear of injuring the larvae and as diffusion is slow across the
surface of an unagitated liquid, it seems probable that the oxygen
tension of the NaCl solution and the sea water in the control
never reached 320 mm. during the experiment.

TABLE III.

EXPERIMENT 2.
Over 50 larvae examined in each case.

A. Examined I Hour after Return to Sea Water.

Time of Exposure.

Solution.

i Hour.

2 Hours.

3 Hours.

Isotonic NaCl at at-
mospheric oxygen
tension

All larvae show

No motion

No motion

bendings

Isotonic NaCl at
slightly increased
oxygen tension. . .

All larvae show
bendings

i /2 larva? show
bendings

i /4 larva? show
bendings

Sea water at slightly
increased oxygen
tension

Larvae normal

Normal

Normal

B. Examined 21 Hours after Return to Sea Water.

Isotonic NaCl at at-
mospheric oxygen
tension

No motion

No motion

No motion

Isotonic NaCl at
slightly increased
oxygen tension. . .

Larvae completely
recovered

i / 1 o larva? show
bendings

Occasional larva
shows bendings

Sea water at slightly
increased oxygen
tension .

Normal

Normal

Normal

VARIATIONS IN OXYGEN TENSION.

139

TABLE IV.

EXPERIMEXT 2.

Over 50 larvae examined in each case.

A. Examined i Hour after Return to Sea Water.

Time of Exposure.

Solution,

i Hour.

2 Hours.

3 Hours.

Isotonic NaCl at at-
mospheric oxygen
tension

1/2 larvae show

No motion

No motion

bendings

Isotonic NaCl at
slightly increased
oxygen tension. . .

All larva? show
bendings

1/20 larva? show
bendings

i/io larvae show
bendings

Sea water at slightly
increased oxygen
tension .

Larvae normal

Normal

Normal

B. Same Larrcc Examined u Hours after Return to Sen U'atcr.

Isotonic NaCl at at-
mospheric oxygen

tension 1/50 larvae show

bendings

No motion

No motion

Isotonic NaCl at
slightly increased

oxygen tension. . . 1/4 larvae show 1/12 larvae show i/S larvae show
bendings bendings bendings

Sea water at slightly
increased oxygen
tension .

Normal

Normal

Normal

The toxicity is certainly markedly diminished by the slight
increase in oxygen tension. The effect is as marked as when
the tension was decreased.

III. In this experiment the oxygen tension of the chamber was
raised to 230 mm. four hours before the experiment was started.
The isotonic NaCl in the beaker " E" was stirred vigorously at
intervals of fifteen minutes allowing the solution to come to equi-
librium at an oxygen tension of 230 mm. before the larvae were
introduced. The larvae were washed twice with this NaCL at
230 mm. oxygen tension, by pumping a few c.c. out through the
tube " T" and they were then introduced into the beaker " E" as

140

ISAAC STARR, JR.

before. The larvae were therefore suddenly subjected to the
change of oxygen tension plus the NaCl.

TABLE V.

EXPERIMENT 3.
Over 50 larvae examined in each case.

A. Observed i Hour after Return to Sea Water.

Solution.

Time of Exposure.

i Hour.

2 Hours.

3 Hours.

4 Hours.

Isotonic NaCl at
atmospheric
oxygen tension
158 mm.

i/io larvae show
bendings

1/5 larvae show
bendings

1/5 larvae show
bendings

No motion

Isotonic NaCl at
approx. 230
mm. oxygen
tension .

All larvae show
bendings

9/10 larvae show
bendings

1/2 larvae show
bendings

1/50 larvae
show
bendings

Sea water at ap-
prox. 230 mm.
oxygen tension.

Larvae normal

Normal

Normal

Normal

B. Observed 18 Hours after Return to S-ea Water.

Isotonic NaCl at
158 mm. oxy-
gen tension ....

1/7 larvae show
bendings

i/ioo larvae show
bendings

1/50 larvae show
bendings

i /i oo lar-
vae show
bendings

Isotonic NaCl at
approx. 230
mm. oxygen
tension ...

9/10 larvae show

i /2 larvae show

3/50 larvae show

2/50 larva?

bendings

bendings

bendings

show
bendings

Sea water at ap-
prox. 230 mm.
oxygen tension.

Normal

Normal

Normal

Normal

This experiment shows the same diminution of toxicity due to
slightly increased oxygen tension.

IV. In this experiment the oxygen tension was raised to 275
mm. before the larvae were introduced, as in the preceding ex-
periment.

VARIATIONS IN OXYGEN TENSION.

141

TABLE VI.

EXPERIMENT 4.
Over 50 larvae examined in each case.

A. Observed i Hour after Return to Sea Water.

Time of Exposure.

Solution.

i Hour.

2 Hours.

3 Hours.

Isotonic NaCl at at-
mospheric oxygen
tension 160 mm. .

i /4 larvae show
bendings

1/50 larvae show
bendings

No motion

Isotonic NaCl at
approx. 275 mm.
oxygen tension. . .

i /4 larvae show
bendings

i/ioo larvae show
bendings

No motion

Sea water at approx.
275 mm. oxygen
tension

Normal

Normal

Normal

B. When observed 24 hours after return to sea water there was no motion
in any case except the control in sea water at 275 mm. which was normal.

The parallelism with the control is striking. The toxic and
antitoxic effects of the increase of oxygen tension must balance at
this tension.

V. In the last experiment the larvae were exposed to an oxygen
tension of 756 mm. (saturated). A stream of oxygen was
bubbled through a flask containing the isotonic NaCl for fifteen
minutes. The larvae were washed twice with this solution, then
were placed in it and oxygen bubbled through for five minutes
more and the flask closed. When some of the larvae were re-
moved at hourly intervals for return to sea water, oxygen was
bubbled through the remaining solution for one minute immedi-
ately after.

TABLE VII

EXPERIMENT 5.
Over 50 larvae examined in each case.

A. Observed I Hour after Return to Sea Water.

Solution.

Time of Exposure.

i Hour.

2 Hours,

3 Hours.

Isotonic NaCl at at-
mospheric oxygen
tension 160 mm.. 3/4 larvae
bendings

show

No motion

1/20 larvae show
bendings

142

ISAAC STARR, JR.

Isotonic NaCl at
approx. 756 mm.
oxygen tension. . .

No motion

No motion

No motion

Sea water at approx.
756 mm. oxygen
tension

Normal

Normal

Heliotropism lost,

otherwise nor-
mal.

B. Observed 18 Hours after Return to Sea Water.

Isotonic NaCl at at-
mospheric oxygen
tension.

1/2 larvae show
bend ings

1/50 larvae show 2/50 larvae show
bendings bendings

Isotonic NaCl at
approx. 756 mm.
oxygen tension. . .

1/50 larvffi show
bendings

No motion

No motion

Sea water at approx.
756 mm

Normal

Normal

Some heliotropic,
some not helio-
tropic. all other-
wise normal.

The increase in toxicity caused by the excessive oxygen tension
is very marked.

The writer wishes to express his thanks to Professor R. S.
Lillie under whose direction the problem was undertaken.

CONCLUSIONS.

1. The removal of most of the oxygen from the solution
markedly diminishes the toxicity of sodium chloride isotonic to
sea water to Arcnlcola larva;.

2. Slight increases in oxygen tension also markedly diminish
the toxicity of isotonic sodium chloride.

3. Saturation with oxygen markedly increases the toxicity of
this solution.

4. At a tension of approximately 2/5 mm. the toxic and anti-
toxic effects balance.

AUTHOR S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, FEBRUARY 28, IQ2I.

TRANSPLANTATION AND INDIVIDUALITY.1

LEO LOEB.2

INTRODUCTION.

In this paper I shall report very briefly on a series of experi-
ments which have been carried out in our laboratory by ourselves
and our associates and in which the method of transplantation
was made use of in the analysis of individuality.

Reactions of tissues serve as indicators with which to judge
the interactions of individualities when parts of organisms are
transferred into a new soil. Conversely, by means of the re-
actions of tissues, which we observe under those conditions, we
attempt to obtain an insight into the forces which are active be-
tween tissues and into their finer biochemical correlations. Upon
these forces depends the preservation of the structure of the
organism and thus its maintenance. Thus we hope to contribute
to the building up of a physiology of tissue in contradistinction
to the physiology of organs. In addition we have studied the lit-
erature of transplantation, in order to obtain data which permit
the comparison of the interaction of individualities in lower and
higher animals and thus to determine whether there are indica-
tions that a gradual change has taken place in the course of evo-
lution. We shall include in our consideration fertilization which
can be considered as an intracellular transplantation.

Transplantation is the separation of a piece of tissue from its
normal surroundings and its transfer into a new environment,
either at a new place in the same host ; this we call " autotrans-
plantation " ; or into a related individual : " syngenesiotransplan-
tation " ; or into a not related individual : " homoiotransplanta-

1 A lecture delivered at the Marine Biological Laboratory, Woods Hole,
on August 3, 1920. A few minor additions to the manuscript have been made
subsequently.

2 From the Department of Comparative Pathology, Washington LJniversity
School of Medicine, St. Louis.

143

144 LEO LOEB.

tion." Within the same species differences of groups and strains,
varieties may arise.

Transplantation into a different species we call " heterotrans-
plantation." Here, nearly related species, which interbreed,
further distantly related species, which do not interbreed, species
which belong to different genera, families, classes can be distin-
guished. To this spectrum of relationships correspond?, with
certain limitations, a spectrum of interactions of tissues after
transplantation.

We judge individuality in man primarily by characteristic
social-psychical reactions. It is, therefore, I presume, originally
a term connoting psychical attributes. Each individual is sup-
posed to form an indivisible whole. It is one organism separate
and distinct from all others. No two organisms can be identical.
This finds the sharpest expression- in the conception of a soul
which is supposed to be the real bearer of the individuality, which
is unlike any other soul and indivisible.

To the psychical individuality corresponds individuality of the
body. The body also is supposed to represent one indivisible
whole, different from every other body ; in short one common
factor uniting all the parts, and distinguishing them from all the
parts of other individuals.

This conception, however, does not quite harmonize with the
mosaic such as has been revealed by Mendelian analysis of in-
heritable characteristics of organisms. Mendelian analysis has
shown an organism to consist of a very great number of unit
factors, the various unit factors being approximately the same
in many organisms of the same species, and the individuals dif-
fering from each other in the mosaic of unit factors of which
each one is composed. A common factor or set of factors present
in all parts of the organism and truly representing its individ-
uality is not provided for in this scheme.

INDIVIDUALITY DIFFERENTIAL.

And yet, such a characteristic present in all parts does exist.
It is the same everywhere in the same organism and differs in
different organisms. We may call this characteristic individual-
ity differential as far as it distinguishes individuals, and species

TRANSPLANTATION AND INDIVIDUALITY. 145

differential as far as in addition it differentiates species, the one
differential being superimposed upon the other. That such a
common factor truly representing the individuality does exist can
be shown through a comparative study of transplantation of one
part of an organism into the same and other individuals, related
and not related, of the same species, and into individuals of
various different species. Such transplantations reveal the pres-
ence of an individuality differential, either directly through the
interaction of tissues and of tissues and body fluids, or in certain
cases indirectly through the immunity which follows such trans-
plantation of tissues or parfs of organisms.

A systematic use of transplantation for the analysis of indi-
viduality is of rather recent origin. Carried on, however, in a
more or less haphazard way the study of transplantation dates
back a considerable number of years. My own interest in trans-
plantation as a means of analyzing the biochemical difference in
the constitution of individuals was first aroused about nineteen
years ago, and again a few years later, when I compared the re-
sult of transplantation of tumors in the same individual and in
another individual of the same species. The transplanted pieces
behaved quite differently in both of these cases.

AUTO AND HOMOIOTRANSPLANTATION.

It is, however, only recently that the difference in the result of
auto and homoiotransplantation has been more generally recog-
nized. On the other hand, that heterotransplantation does usu-
ally not succeed has been established considerably earlier.

THE MECHANISM WHICH DETERMINES THE INTERACTION OF

TISSUES.

For a number of years my associates and myself have carried
out experiments tending to analyze the factors which connect
individuality and transplantation, and here I wish to describe
very briefly a few of our experiments and to draw some more
general conclusions, the latter merely in a tentative manner. Our
experimental analysis is not yet concluded and at the present
time there are under way certain investigations which we hope
will help to clear up some of the doubtful points. For our pur-

146 LEO LOEB.

pose it might be best to select a few examples of our experiments
in order to illustrate some of the more important factors that
come into play under various conditions of transplantation. We
shall first discuss transplantation of pigmented skin, in the guinea
pig, then as examples of glandular organs, the thyroid gland and
kidney, and lastly the uterus as an organ containing a variety of
tissues, epithelium, myxoid or predeciduomatous connective tissue
and unstriated muscle.

TRANSPLANTATION OF PIGMENTED SKIN IN THE GUINEA PIG.

If we transplant black skin of the guinea pig into a defect in
white skin of the same individual, the black skin usually heals in
and even begins after a few weeks to penetrate into the neigh-
boring white skin. If instead we transplant it into a defect in
the white skin of another guinea pig, the skin may temporarily
heal in, but sooner or later it is cast off, in the majority of cases
at an early date following the grafting. In a few cases, however,
the transplant took. I suspect these were cases in which host and
donor were related to each other, so that in realitv we had to

j

deal with syngensio- rather than homoiotransplantation. Now
in these exceptional cases the black skin did not only not pene-
trate into the neighboring white skin, but on the contrary, it
gradually became paler and its pigmentation disappeared in the
end entirely. The black skin became transformed into white skin.
Microscopically we found in such cases in addition a gradual ac-
cumulation of lymphocytes under and in the strange epidermis.
The number of these cells need, however, not be very marked.
Lymphocytes are small cells with a relatively prominent round
nucleus which originate in lymph glands, gain entrance into the
circulation and are capable of active movement, which, how-
ever, is probably not as active as in the ordinary polynuclear leu-
cocytes. These lymphocytes, as we stated, are attracted by the
foreign skin, but not by the skin of the same organism. We
notice after this kind of homoiotransplantation, a primary in-
compatibility between the strange skin and its environment. As
a result of this incompatibility between body fluid of the host
and the transplant changes take place in the metabolism of the
latter which cause the deficient healing in of the transplant and

TRANSPLANTATION AND INDIVIDUALITY. 147

which later attract the lymphocytes. In addition there occur ab-
normal reactions in the transplant which render impossible the
normal restitution of pigment in the graft. Some of these
changes in the transplant may take place so soon after trans-
plantation that the conclusion suggests itself that in this case an
incompatibility between the graft and the body fluids of the host
interferes directly with those tissue reactions on which depends
the healing in of the graft and the reestablishment of the normal
pigmentation. How far these primary changes in metabolism of
the graft injure the transplant directly and to what extent they
act possibly through induced changes in vascularization of the
graft still remains to be determined. It is, however, more prob-
able that the incompatibility between body fluids and graft leads
to a direct interference with the transplant.

TRANSPLANTATION OF THYROID AND KIDNEY

If we transplant thyroid gland into the subcutaneous tissue
only the peripheral gland tissue survives, the center of the graft
being ill nourished in the first few days following the detachment
of the gland from its soil dies. The peripheral gland acini pro-
liferate soon after transplantation. After autotransplantation
blood and lymph vessels grow through this ring of well-preserved
and temporarily growing thyroid tissue into the necrotic center,
and especially in the peripheral part of the latter they form a ring
of vessels which is very noticeable ; from here vessels penetrate
into the center of the necrotic material. They are accompanied
by a relatively limited number of fibroblasts which at the inner
aspect of the thyroid ring form a loose, almost myxoid connective
tissue around the blood and Imyph vessels, a tissue not unlike that
found in certain stages of development in the embryo. Further
removed from the thyroid ring, in the real center of the necrotic
material, there may be produced a nucleus of dense fibrous tissue.
But its existence is only temporary after autotransplantation.
Sooner or later the blood and Imyph vessels and fibroblasts pene-
trate into it, absorb it and substitute for it the same kind of loose
vascular connective tissue ; this again gives way more and more
to the peripheral real thyroid tissue. As a result of the absorp-
tion of the central connective tissue and of the marked develop-

148 LEO LOEB.

ment of the peripheral thyroid tissue and of the good vascular
supply the normal organ structure is more and more reestab-
lished. This occurs within three to five weeks after transplanta-
tion. Lymphocytes are at no time prominent in the transplant.

After homoiotransplantation of the thyroid gland the first
stages are similar to those observed after autotransplantation ;
but soon two important differences become noticeable. The well
developed ring of lymph and blood vessels in the peripheral part
of this center is much less prominent after homoiotransplanta-
tion ; the connective tissue ingrowth, on the other hand, is much
more marked. The whole center becomes soon converted into a
dense fibrous mass, much larger in volume than in the autotrans-
plant. It may still be well supplied with vessels, but they are less
prominent than after autotransplantation. In contradistinction
to what happens in the autotransplant this fibrous mass is not
absorbed and not substituted first by myxoid and later by thyroid
tissue. This transplant never attains similarity in structure with
the normal gland. Even around the individual acini the fibro-
blasts may become active and form fibrillar and fibrous tissue.
The activity of the connective tissue may go still further, and oc-
casionally fibroblasts may penetrate into the interior of some
thyroid acini and help to destroy them.

The direct destruction of the graft by the host is a very character-
istic feature after homoiotransplantation. It is however, not so
much the activity of the fibroblasts as of the Imyphocytes which
brings about this result. While in the autotransplant the lympho-
cytes are practically absent, in the homoiotransplant they begin
to appear as early as five and seven days after transplantation.
However, they become more prominent only about nine or ten
days after transplantation, and from then on they rapidly in-
crease in number in many cases. They approach the transplant
mainly by way of the lymph vessels, and in consequence in the
homoiotransplant the lymph vessels may become as prominent as
if they had been injected. This I found especially noticeable in
the rat, on account of the peculiar distribution of the lymph ves-
sels at the inner aspect of the thyroid ring. The lymphocytes
appear at this place in especially large masses and from here
they penetrate in a peripheral direction into the thyroid acini

TRANSPLANTATION AND INDIVIDUALITY. 149

proper and destroy them. Other lymphocytes approach the
thyroid from the periphery and from here push forward in a
central direction. All these lymphocytes surround, invade and
destroy the thyroid tissue directly after homoiotransplantation
and they are, therefore, the chief agent of destruction. They
interfere with the nourishment of the acini by surrounding them
and cutting them off from the vessels ; they exert furthermore a
direct pressure upon them and invade and substitute them. In
addition the fibrous tissue which is so prominent after homoio-
transplantation also tends to shut off the nourishing material
from the vessels. Furthermore it compresses the glandular struc-
tures and fibroblasts occasionally invade them, as we have stated
above. These lymphocytes and fibroblasts form, therefore, an
agency of attack which usually succeeds in destroying the homoio-
transplant of the thyroid somewhere between the fifteenth and
twenty-eighth day following transplantation.

We may again assume that in the strange host the body fluids
are not completely adapted to the transplanted gland structures.
In consequence their metabolism is to a certain extent altered.
This alteration, however, is not of sufficient intensity to cause a
direct destruction of the transplanted glandular structures. This
alteration in metabolism attracts the lymphocytes, diminishes the
vascularization of the transplant and increases the invasion of
the graft by fibroblasts, which, however, under these altered con-
ditions do not remain intact, succulent cells, but instead form
fibrillar and dense fibrous tissue.

In principle it is similar after transplantation of the kidney ;
but the greater denseness of this material brings about certain
minor modifications in the result. Thus in the kidney after auto-
transplantation we do not find the inner ring of lymph and blood
vessels so noticeable as in thyroid autotransplant. But in all
essential respects the conditions are parallel to those after trans-
plantation of thyroid. The formation of fibrous tissue is much
more prominent after homoiotransplantation of the kidney than
after autotransplantation. The lymphocytes play here a similar
part. In the end the homoiotransplant is again destroyed in the
same way as in the case of the homoiotransplanted thyroid. But
this destruction of the kidney may be completed at a somewhat

I5O LEO LOEB.

later date. After autotransplantation kidney tissue usually re-
mains preserved.

We have begun comparative studies of auto and homoiotrans-
plantation in other species. There are indications that in prin-
ciple conditions are similar in those species which we have begun
to study, but there may exist some quantitative and perhaps also
some qualitative differences in different species, just as we found
certain differences in the behavior of the thyroid and kidney.
Thus to mention only one difference, we found that in the rabbit
the polynuclear leucocytes are much more prominent after trans-
plantation than in either rat or guinea pig.

SYNGENESIOTRANSPLANTATION.

If instead of transplanting the thyroid gland in the guinea pig
into not related individuals of the same species, we transplant it
into nearly related individuals of the same family, for instance
from brother to brother or sister, or from parents to children, we
find an intermediate condition between the results of auto and
homoiotransplantation, or rather we find all degrees of an inter-
mediate condition. Thus while after autotransplantation lympho-
cytes are almost absent and after homoiotransplantation they
usually begin to become prominent during the latter part of the
second week, they may after syngenesiotransplantation, appear
as late as the second half of the fourth week ; sometimes- they
may appear somewhat earlier or at other times later. But they
usually end by destroying the transplant even in relatives. We
find therefore a noticeable delay in the reaction on the part of the
host. Evidently at first the metabolism of the syngenesiotrans-
plant is little altered by the body fluids of the host. Gradually,
however, the alteration becomes here also sufficiently strong to
attract the lymphocytes. As we have stated above the vascular
and connective tissue reaction on the part of the host tissue usu-
ally takes place between the seventh and twelfth days after trans-
plantation, at a time, therefore, when the alteration of the metab-
olism has usually not yet become very marked. The blood and
lymph vessel supply may therefore in the case of svngenesio-
transplantation be as satisfactory as after autotransplantation, al-
though later the lymphocytes begin their destructive work. There

TRANSPLANTATION AND INDIVIDUALITY. 151

is a prolonged latent period in the case of syngenesiotransplanta-
tion and the vascular and connective tissue reaction falls into this
latent period.

As we have stated above we find in syngenesiotransplantation
all degrees of intermediate condition in different individuals of
the same family, for instance in different brothers. In some indi-
viduals host and graft may be almost as unsuitable to each other
as after homoiotransplantation. In such a case the metabolic
alteration of the transplant may have reached a considerable
strength at the period, when the vascular and connective tissue
reaction is determined. In this case the dense fibrous tissue of
homoiotransplantation may be produced in syngenesiotransplan-
tation.

Most beneficial seems to be the exchange of tissues between
brothers and sisters. Very good, but perhaps sligthly less advan-
tageous is the transplantation from parents to children. Pecul-
iarly enough, we found that after transplantation of the thyroid
in the guinea pig from child to mother the result is about as un-
favorable as after transplantation into a totally strange individual
of the same species. This peculiar phenomenon we intend to
analyze still further. A similar intermediate condition we found
after multiple simultaneous transplantation of organs into rela-
tives in the rat. In this case we have, of course, to take into con-
sideration the fact that different organs and parts of organs differ
in their power of resistance in the period following transplanta-
tion, and that the relation between the parenchyma and stroma
also differ in different organs. Again we found in the latter
series all degrees of intermediate condition and different organs
behaved in the same host in an analogous way, if they, came from
the same donor.

TRANSPLANTATION OF UTERUS.

In the transplantation of the uterus, we are especially interested
in the behavior of the myxoid or predeciduomatous, cellular con-
nective tissue and in the unstriated muscle tissue which are char-
acteristic of the uterine structure. Soon after transplantation the
death of a part of the myxoid connective tissue, which is not un-
like connective tissue present in the embryo, and of the unstriated

152 LEO LOEB.

muscle takes place in the auto as well as in homoio graft. This is
the result of the injury caused by the process of transplantation
and the defective nourishment directly following the grafting.
But from about the sixth to the tenth day following transplan-
tation a much better, more complete recovery of the myxoid and
unstriated muscle tissue takes place in the auto- than in the homo-
iotransplant, and while subsequently in the autotransplant these
two tissues maintain themselves, in the homoiotransplant a grad-
ual substitution of the injured myxoid and muscle tissue by
fibrous tissue takes place sometime between the fourteenth and
twenty-fourth day. On this soil of fibrous tissue the epithelium
does not thrive so well as on the myxoid connective tissue of the
autotransplant ; it decreases therefore in size, and becomes lower ;
later it is again attacked by lymphocytes and thus gradually de-
stroyed. The lymphocytes though appear here somewhat later
than in the homoiotransplanted thyroid and kidney ; but as in the
case of the other organs their attack is mainly directed against
the epithelial tissue, although they do not leave entirely free some
of the other tissues. Again we find as the result of those ab-
normal substances which form after homoiotransplantation, — we
may call these substances homoiotoxius — an attraction of lympho-
cytes and an altered behavior of connective tissue cells. But in
addition we find in this case a very early injurious influence of
the hcmoiotoxins on two sensitive tissues, namely myxoid or pre-
deciduomatous and unstriated muscle tissue. They show the in-
jurious effect of the conditions prevailing in the strange host at
a time, when lymphocytes have not yet had a chance to play any
considerable part. How far these two tissues are injured directly
by an inadequate constitution of the body fluids of the host (by
the " homoiotoxins "), and how far the latter influence primarily
the vascular supply, the latter in turn influencing the life of the
myxoid and unstriated muscle tissue, is uncertain at present. It
is however more probable that their destruction is primarily due
to the inadequacy of the body fluids rather than to an insufficient
vascular supply, which latter would then be the direct conse-
quence of the lack of adaptation between body fluids and trans-
planted tissue. There is another point of interest in the trans-
plantation of the uterus. We find that the primary transforma-

TRANSPLANTATION AND INDIVIDUALITY. 153

tion of myoxid into fibrous connective tissue leads secondarily to
changes in the epithelium which rests on the connective tissue.
We find thus a chain reaction and in addition to the direct results
indirect results of the homoiotoxin action.

MECHANISM OF THE REACTIONS.

These experiments prove that the introduction of parts of
organs or tissues which originated in a strange individual causes
disturbances which lead to changes similar to those found as the
result of the action of toxic substances. These substances act not
unlike those given off by certain microorganisms, as for instance
the tubercle bacillus, which cause changes of a not acute char-
acter. Lymphocytes are attracted and besides the relations be-
tween various tissues are quite markedly altered. We have every
reason to assume that these disturbances are due to products of
metabolism given off by the introduced tissue, which act as homo-
iotoxins. We have learned that the action of these substances
is graded in accordance with the relationship between donor and
host. It is as yet doubtful, how far these disturbing substances
are those given off in the normal metabolism of the transplanted
cells — substances which are toxic merely because they act on a
strange host — and how far they are the product of an abnormal
metabolism of the introduced cells, the pathological change being
due to the action of the body fluids of the new host upon the
strange cells. It is probable that the second alternative holds
good at least in many cases. Landsteiner, von Dungern and others
have shown that in man certain groups of individuals can be dis-
tinguished according to the interaction of blood cells on the one
hand, and agglutinins preformed in the blood on the other hand.

While such agglutinins have not been observed in animals, in
certain cases it has been found possible through immunization
with blood corpuscles belonging to the same species, but to dif-
ferent individuals to produce hemolysins which dissolve cor-
puscles of the same species and combine especially with the cor-
puscles of the individual whose blood had been used for injection.
These observations, as well as our own experiments to which we
referred already, as wrell as others to be mentioned later render it
at least probable that such an interaction takes place between a

154 LEO LOEB.

constituent of the body fluids of the host and the strange cells and
that this interaction leads to the formation of toxic substances.
While these toxic substances probably interfere in an injurious
manner with certain sensitive tissues and lead to their destruction
in an indirect manner, in the case especially of certain epithelial
structures they merely alter the metabolism in such a way as
would be perfectly compatible with the life of the tissues. But
this alteration in metabolism sets into motion secondary processes,
in consequence of which lymphocytes, vessels and fibroblasts
show an altered relation to the transplanted epithelium and these
tissue reactions lead secondarily to the death of the transplant.

SYNGENESIO AND HOMOIO TOXINS AS PRODUCTS OF METABOLISM.

We have assumed that the substances which are characteristic
of the individual and which call forth the reactions which we
have described, pass into the surrounding medium as the result of
the metabolism of the tissue. In all probability they are not
merely decomposition products of proteins. In order to exclude
this latter interpretation we studied the behavior of the host tissue
towards homoiotransplantecl bloodclots in which the blood cells
live for a certain period without, however, carrying on an active
metabolism. We found that such blood cells do not call forth
any of the reactions characteristic of homoiografts ; neither
lymphocytes nor connective tissue nor vessels behave towards
them in any specific way whatever. These homoiodifferentials
are as far as we know, common to all the active tissues of an
organism. They are the same in the different organs of the same
animal. As we stated above, we could show this fact in the rat
by multiple simultaneous transplantation of different organs of an
individual into the same host. Under this condition all the organs
elicit proportionately the same individuality reaction, because the
relation between the individuality differentials of host and donor
is everywhere the same. Particularly the lymphocytic reaction
allows us to estimate this relationship of the individuality differ-
entials in an approximately quantitative way, as our syngenesio-
transplantations have shown.

TRANSPLANTATION AND INDIVIDUALITY. 155

MULTIPLE AND SUCCESSIVE TRANSPLANTATIONS.

We can demonstrate this fact also by multiple simultaneous
transplantations of the same kind of organ, for instance, the
thyroid from different individuals into the same host. The lobes
of thyroid taken from the same animal behave then in an approxi-
mately similar manner, while the lobes taken from different ani-
mals may behave very differently, each calling forth a lymphocytic
reaction in a quantitative way in accordance with the relationship
between host and donor. One piece can call forth a marked reac-
tion, while at the same time and in the same host another piece
proves rather indifferent. This indicates that the reaction is of
an entirely local character, called forth by the substances diffusing
from the transplanted cells into the neighboring tissue. It is not
primarily a general reaction of the character of an immune -
reaction, which would depend mainly on the presence of sub-
stances originating in response to the inoculation of the tissue and
carried through the circulation equally to all parts of the body.
The local character of the reaction we could prove still in another
way, namely by carrying out successive transplantations of the
same kind of organ into the same host. Under these conditions
the latent period of the lymphocytic and connective tissue reac-
tion was approximately the same after the first and second trans-
plantation. An immune substance which could have accelerated
the second reaction had, therefore, not been formed. In certain
cases, however, such an immune substance may actually develop
and hasten the appearance of a reaction around the transplant.
This takes place after inoculation with certain tumors. Thus it
comes about that the function of the lymphocytes has been misin-
terpreted in the case of tumor inoculations. It is believed that
they are solely concerned in the production of an immunity
against tumor growth. Our observations on the action of lymph-
ocytes in the case of transplantation of ordinary tissues, which
date back a considerable number of years, clearly prove that the
role of lymphocytes is a much more general one, namely, a direct
and local, quantitatively graded response to the homoio and syn-
genesiotoxins. The immunity reaction in certain cases of tumor
transplantation is merely a special case in a set of phenomena of
much wider biological significance.

156 LEO LOEB.

CONSTANCY OF THE INDIVIDUALITY DIFFERENTIAL.

This individuality differential is a relatively constant factor that
varies not at all or at least to a very limited extent in the same
individual after the animal has reached the age, when it is able to
obtain its nourishment independently of its mother. Such condi-
tions as pregnancy, temporary undernourishment, certain infec-
tions do not suspend the function of the individuality differential.
In addition we could show in a separate series of experiments
that particularly those changes which call -^orth compensatorv
hypertrophy in the thyroid gland do not noticeably interfere with
the reaction against the homoiodifferential. The lymphocytes
may invade in enormous numbers the gland even after it has
become hypertrophic and again in the end they succeed in de-
stroying it.

THE EFFECT OF HOMOIOTOXINS ON OTHER GROWTH PROCESSES.

While thus compensatory hypertrophy does not noticeably
modify the action of the homoiotoxins, the homoiotoxins may
on the other hand, as our recent experiments have shown, to
some extent interfere with the development of compensatory
hypertrophy of the thyroid gland, not only by bringing about the
destruction of the gland, but apparently also by diminishing the
frequency of the hypertrophic changes. In a similar manner
our previous experiments had shown that homoiotoxins may to
some extent interfere with the production of the maternal placenta
and the placentomata such as they can be produced experimentally
in the normal as well as in the homoiotransplanted uterus. After
homoio-transplantation the experimental formation of placenta is
diminished. The homoiotoxins interfere, therefore, to some ex-
tent not only with purely regulative processes such as occur after
transplantation of organs, but also with those changes which lead
to compensatory hypertrophy and to placenta formation. They
evidently have an injurious influence on a variety of growth
processes, and diminish the intensity of those tissue reactions
which initiate placenta formation.

TRANSPLANTATION AND INDIVIDUALITY. 157

HETEROTRANSPLANTATION.

If we compare with these results of auto, syngenesio and
homoiotransplantation, heterotransplantation (transplantation
into a strange species ) we find some interesting differences.
After heterotransplantation tissues generally live only a short
time, which varies between a few days or even less and two weeks
or slightly more. In a few exceptional cases tissue may even live
as long as three or four weeks. For a short period there may be
found a slight proliferation of the heterotransplant. But usually
the injurious action of the host and particularly of its body-
fluids is very marked after heterotransplantation. The quantity
of living, well preserved tissue is therefore very much reduced.
These differences between the hetero and homoiotransplant are
usually quite marked as early as the latter part of the first and
the beginning of the second week. There may be added to the
direct injurious effect of the host and its body fluids a destructive
action of lymphocytes and an invasive action of fibroblasts. But
both of these are usually relatively slight ; and especially the
lymphocytic reaction is markedly less prominent after hetero-
than after homoiotransplantation. Fibroblasts and bloodvessels
of the host show little activity around the heterotransplanted
parenchyma. The vascularization is therefore poor and the
number of fibroblasts growing directly around the heterotissue is
usually restricted. The connective tissue that does grow has the
tendency to form fibrous tissue. At some distance from the
transplanted parenchmya the fibroblasts and bloodvessels behave
otherwise as if they had to deal with an inert foreign body. While
thus the lymphocyte and connective tissue contribute only slightly
to the destruction of the heterotransplant — the fibrous tissue ex-
erting an injurious pressure on the heterotransplant — large masses
of lymphocytes may collect in the tissue surrounding the hetero-
graft. These lymphocytes, however, are relatively innocuous.

If we ask, how it comes about that after heterotransplantation,
notwithstanding the greater strangeness of host and donor, the
destructive action of the lymphocytes is not only not more marked
than after homoiotransplantation, but on the contrary much more
restricted, we may suggest that after heterotransplantation the

158 LEO LOEB.

grafted tissue is injured to so marked an extent, that a depression
in metabolism occurs and the quantity of toxic substances attract-
ing the lymphocytes and produced in the metabolism of the tissue
is much diminished. All this applies to the transplantation into
not nearly related species. If we transplant into nearly related
species, results are much better as has been established by W.
Schultz. It seems that in this case the tissues behave almost like
homoiotransplants. However it appears probable to me that a
comparative study — which so far has not yet been made,— would
show that even in this case the results are distinctly less good
than after homoiotransplantation. If we disregard the hetero-
transplantations into nearly related species, there is after hetero-
transplantations not a close connection between the results of
transplantations and the relationship of the species used. The
heterotransplanted tissues are all so near the minimal threshold
which just permits life for a short time, that various secondary
factor often become of more importance in determining the
duration of life and extent of proliferation of the graft than the
character of the species differentials. There exists, however, as
we have shown an indication that even here such a relationship
enters as one of the determining factors.

SPECTRUM OF RELATIONSHIPS AND INTERACTION OF TISSUES.

If we now compare the effect of the various kinds of trans-
plantations on the character of the interactions between the tissues
of host and graft, we come to the following conclusions :

i. The autotransplants have the greatest degree of efficiency in
preserving the integrity of the graft and in maintaining inviolate
the boundaries of the organs and in preventing the ingrowth of
the connective tissue of the host. The autotransplant behaves in
this respect most like the normal organ. From there a decrease
in efficiency takes place if we pass to the syngenesio- and to the
homoiotransplants. In the heterotransplant this power of pre-
serving the integrity of the graft and of warding off the attack by
strange connective tissue has become very slight. There is, how-
ever, still noticeable a slight action of the transplanted paren-
chyma even in this case ; but on the whole the tissue behaves not

TRANSPLANTATION AND INDIVIDUALITY. 159

unlike an inert foreign body. Correspondingly the amount of
fibrous tissue formed increases and the extent of vascularization
decreases in the direction from atitotransplant to heterotransplant.
The ability of the transplanted parenchyma to maintain a domi-
nance over the stroma, to keep the fibroblasts intact, and if pos-
sible in a myxoid condition and to cause the absorption of the
connective tissue stroma decreases in the direction from auto to
heterotransplant. Similar is the curve which represents the
stimulating power of the parenchyma on the vesselgrovvth. It
needs further investigation to determine how far the vascular re-
action is a primary phenomenon and how far the connective tissue
reaction depends upon it. The lymphocytic reaction on the other
hand reaches a maximum in the homoiotransplant and decreases
again in the heterotransplant.

THEORETICAL CONSIDERATIONS. CONTACT SUBSTANCES, AUTO-
SUBSTANCES AND TOXINS.

— - •««

Our results are best understandable, if we assume that the
cells in their metabolism give off a series of substances which
regulate the relation with certain other kinds of cells, and in par-
ticular the relation of parenchyma (epithelial, glandular, special
kinds of connective tissue cells and unstriated muscle) to the
surrounding connective tissue, blood and lymph vessels and
lymphocytes.

We must conclude that the cells living under what we might
call " autocondition," give off another substance or another set of
substances from those living under " syngenesio," "homoio,"
"hetero" conditions. We have found a graded injuriousness of
the latter kind of substances for the cells of the surrounding
tissues. They stimulate the surrounding tissues to reactions
which are pathological and which in the end lead to the destruc-
tion of the cells which produce these substances and call forth
these reactions. We have therefore called these substances " syn-
genesio," " homoio," " hetero " toxins. Conversely we may des-
ignate those substances which are given off by the normally func-
tioning cells under " auto " surroundings as " auto " substances.
They regulate the action of connective tissue cells, vessels and

I6O LEO LOEB.

lymphocytes in the most adequate way, and in a manner peculiar
to each organ, in a way which keeps away lymphocytes, which
limits the activity of fibroblasts which under those conditions
cannot invade the tissue of another kind and merely enter in such
relations with these tissues as are demanded by the normal struc-
ture and function of the organ. The auto substances thus bring
about the restitution of the normal structure in an at first disor-
ganized organ, just as these substances maintain the normal struc-
ture and function in the normal organ. The syngenesio, homoio
and hetero conditions on the other hand lead to those pathological
conditions which we have described. There are other facts which
equally point to the conclusion that substances wrhich in general
we may designate as " contact substances " are given off and
regulate the relations of various kinds of tissue to each other.

We may mention a few examples of conditions under which
contact substances come into play.

(a) In our study of the cyclic changes in the mammary gland
we found a relation between the state of the glandular paren-
chyma and the surrounding connective tissue stroma. We found
that an active gland has an active stroma, while a resting, retro-
gressing gland has a resting, more or less fibrous stroma.

(6) Similarly we find around the various gland ducts which
are metabolically inactive usually a resting fibrous stroma, in con-
tradistinction to the cellular stroma around the active gland tissue.

i c ) Wherever cells of the parenchyma are multiplying we
generally find the stroma cells and the vessels to become like-
wise active and conversely where connective tissue cells and the
vessels are active the cells of the parenchyma, for instance, epi-
thelium, receive a stimulus to grow.

(d) There are indications that contact substances play also
a certain role during embryonic life and that here they help to
determine the formation of some organs. Thus a lens producing
contact substance is given off by the optic disc. It stimulates
growth and a special differentiation in the overlying ectoderm.

Thus we know at the present time a large class of contact sub-
stances and it is probable that future studies will still add to the
number of these substances.1

1 Several years ago Dr. Walsh and the writer found that a substance given
off by the ovum determines the development of the follicle in the ovary.

TRANSPLANTATION AND INDIVIDUALITY. l6l

CONTACT SUBSTANCES AND HORMONES; THEIR RELATION TO THE
INDIVIDUALITY DIFFERENTIAL.

These contact substances are contrasted with the hormones,
for instance those given off by the corpus luteum which regulates
the activity of the mucosa of the uterus and probably of certain
other organs. They are given off by certain organs, and carried
through the circulation to distant organs. They have a specific
distant action. In contradistinction to some of the contact sub-
stances these distance substances do not possess an individuality
or even a species differential.

CONTACT SUBSTANCES AND CHANGES IN OLD AGE.

I cannot conclude a consideration of this aspect of the subject
without pointing out the similarity of conditions which we find
under " homoio conditions " with those found in old age. In both
cases we observe a tendency to the formation of a fibrous stroma
and to a decrease in good vascularization. This condition also
corresponds to that found in states of metabolic inactivity. May
we not therefore refer old age changes not only to the lack of
certain hormones given off by glands (as for instance thyroid and
corpus luteum) with internal secretions, but also to a quantita-
tive or sometimes perhaps even to a qualitative change in the
character of contact substances, to a diminution in the produc-
tion of what we have called the " auto substances " ? Such a
diminution in auto substances would be the necessary result of a
diminished activity of the parenchyma. Thus a vicious circle
would be established. The diminished activity of the paren-
chyma causes changes in stroma and vascularization and the latter
further depresses the activity of the parenchyma.

INHERITANCE OF THE INDIVIDUALITY DIFFERENTIAL.
We have seen that there is in each individual and in each cell,
at least in the large majority of all cells of an indivdual, an indi-
viduality differential which is present in addition to the ordinary
Mendelian unit factors. The inheritance of the latter has given
rise to numerous investigations which on the whole have tended
to show the general applicability of Mendelian rules in the in-

1 62 LEO LOEB.

terpretation of phenomena of inheritance. The concept of the
individuality differential is too new to have received much at-
tention from students of heredity. But von Dungern studied the
group agglutinins to which we have referred above and mentions
that they are inherited according to the rules of the inheritance
of simple Mendelian monohybrid factors. G. Schoene likewise
inquired how the characteristic of tolerance for skin grafts was
transmitted from parents to offspring and this author also came
to the conclusion that it was inherited according to the rules of
simple Mendelian heredity. Neither of these investigators, how-
ever, gives convincing data in this respect nor does Schoene refer
to the finer tissue reactions in his interpretations. Cur own ex-
periments lead us to a different conclusion. In the union of a
female, and male germcell of two individuals belonging to the
same species, two different individuality differentials are hybrid-
ized. How are the individuality differentials of the children?
Do they follow that of the father or that of the mother or do
the differentials of some children follow that of the father, while
those of others follow that of the mother? \Ye find that the
individuality differentials of the offspring are not identical — at
least in the large majority of cases — with the individuality differ-
ential of either of the parents, but that they have an intermediate
character; there exists, however, as we have shown, all transi-
tions between one and the other of the two individuality differ-
entials, if we compare the behavior of different children. We
have therefore a mode of intermediate or blending inheritance.
If we interpret this fact in Mendelian conceptions we must con-
clude that the individuality differential is represented by multiple
factors.

INDIVIDUALITY AND SPECIES DIFFERENTIAL OF SUBSTANCES.
SPECIFICALLY ADAPTED SUBSTANCES.

The cells and tissues of an organism contain, as we have seen,
the individuality and species differential ; they characterize each
individual and distinguish it from all other individuals. They
make a real indvidual out of a conglomeration of cells and tissues.
In order that these cells and tissues develop into an individual

TRANSPLANTATION AND INDIVIDUALITY. 163

organism and subsequently function as such, substances are given
off by the tissues and organs which act upon adjoining or distant
parts of the body and thus bring about a correlation of functions
which makes possible the orderly development and maintenance
of the organism. Some of these substances still preserve the
individuality and species differential. We may, therefore, con-
clude that not only cells, but also substances given off by cells
may still have the individuality or species differential. This
applies to some of the contact substances which we have postu-
lated. They may have the individuality differential. In most
substances, however, the presence of an individuality differential
cannot be demonstrated, but merely the existence of a species dif-
ferential. Of especial interest among these are certain sub-
stances which make use of this species differential in their func-
tion and which are therefore most effective if they interact with
other substances having the same species differential or rather
a supplementary species differential adapted to the first differ-
ential. Such substances I have called " specifically adapted sub-
stances.'' To this group of specifically adapted substances belong,
as I found in my earlier work, substances which are present in
the tissues and erythrocytes and which interact with a constituent
of the circulating bodyfluid in order to accelerate the clotting of
the blood and lymph. These substances I have called tissue
coagulins. They seem to be id~entical with the thrombokinases.
The species differential fulfills in this case a definite and im-
portant function. Subsequently Hedin discovered in the gastric
juice a substance inhibiting the milk coagulating enzyme. Both
these substances interact with the species differential. More re-
cently Dr. Frank Lillie found a specificially adapted relation be-
tween a constituent of the egg, an agglutinin, and the spermatozoa
of the same species. In this case, however, the common species
differential functions in substances belonging to two different
individuals, while in the former cases the interaction occurs in
substances of the same individual. On the other hand a large
number of important substances which have the function to cor-
relate and unify the action of various organs, particularly certain
growth substances and the common hormones are not only not

1 64 LEO LOEB.

individual specific but not even species specific. This applies for
instance to the growth substances emanating from the corpus
luteum in which I found the absence of the individuality differ-
ential. The lack of the species differential applies to growth sub-
stances extracted from the placenta, to the growth substances
which determine the formation of the lens, to substances inducing
metamorphosis in amphibia and compensatory hypertrophy in
mammals ; it applies to the common hormones of the adrenal gland
and thyroid.

In this case we have to deal with relatively simple substances,
some of which are apparently of a lipoid nature, while others are
of a still simpler composition. On the other hand we have every
reason to assume that the individuality and species differentials
are proteid substances or at least that they occur only in combina-
tion with proteid substances.

In the case of the tissue coagulins we have found that the spe-
cies differential of the tissues interacts with an adapted substance
in the body fluids. In a similar way we find in general an adap-
tation between body fluids and cells which is based on the pres-
ence of the individuality and spe?ies differentials. It can be
demonstrated directly in the case of the natural hemolysins, but
probably applies as we have pointed out above, to the relations of
all tissues to body fluids. It exists, as we 'found recently, even in
the case of invertebrates where we established the interaction of
species differentials in the case of the experimental amcebocytic
tissue and the blood serum. We shall refer to it again later. It
also applies to the relation between the antigens and immune
substances.

SUPPLEMENTARY DIFFERENTIALS IN BODY FLUIDS.

In all those cases the individuality or species differential of the
tissues interacts with an analogous substance in the body fluids.
It is, however, not certain that the substance or group in the body
fluids is identical with that in the tissues, although they fit into
each other specifically. A difference between the two could be
demonstrated in the case of tumor immunity. While species im-
munity can be produced with body fluids as well as with tissues
of a certain species, individuality immunity can only be produced

TRANSPLANTATION AND INDIVIDUALITY. 165

with the tissue differential, but not with the corresponding dif-
ferential in the body fluids. We may designate the substance
present in the body fluid and interacting with the individuality
and species differential of the tissues as the supplementary indi-
viduality and species differential.

PHYLOGENETIC AND ONTOGENETIC EVOLUTION OF THE INDIVID-
UALITY AND SPECIES DIFFERENTIAL.

The analysis of the individuality differential which we have
given so far applies altogether to the higher animals, mammals
and birds. Almost all the experiments to which we referred in
this paper were made in these two classes. Does this fine differ-
entiation of individuality, the delicate discernment of individual
relationship such as we have found in the cells and tissues of
higher animals, also exist in the lower animals? Is it a charac-
teristic of all animal cells or has it gradually evolved together
with other differentiations in the course of evolution? Have the
differentials, or rather the reactions they call forth, had an evo-
lution like structure and certain functions? So far as I am ac-
quainted with the literature this question has never been put and
no planful investigations have been carried out tending to answer
it. In a general way, however, it is known that in lower forms
and in earlier embryonic stages transplantability is greater, in
correspondence with the greater regenerative power, of these
organisms, or in dependence on the greater power of isolated parts
of these beings to sustain themselves separated from the remnant
of the animal, as particularly W. Schultz has pointed out. But
the evolution of species and individuality differentials has not
been considered in a conscious way as far as I am aware.

I have recently studied the literature of transplantation with
the view of determining whether the experiments which were
carried out by various investigators for the solution of problems
of a different character might throw some light on this question.
While in some respects I found the evidence here and there some-
what contradictory, still I believe that certain conclusions may be
arrived at on this basis with a fair degree of certainty.

To begin with invertebrate embryos, the experiments of
Driesch, Morgan, Goldfarb and others show that parts of em-

166 LEO LOEB.

bryos of the same species can be readily grafted upon each other.
The homoiodifferential does apparently not play any role. These
embryos on the other hand react against union with the part of
an embryo belonging to a different species, against an heterodif-
f erential ; the reaction, however, seems to be less marked than in
mammals. Similar are the results obtained with transplantation
of gonads in larvae of lepidopterse (Meisenheimer, Kopec). In
nearly related species the results of grafting are better than in
distant species.

It is similar in adult invertebrates. Here also no difference
seems to exist between auto- and homoiotransplantation and while
a reaction takes place against heterotransplants, it is again con-
siderably less pronounced than in the higher animals. Of value
are here especialy the experiments of Joest, Harms, Leypoldt
and Rabes in Lumbricidae and of Wetzel in Hydra. While hetero-
differentials and also reactions against heterodifferentials exist
here, the reactions are much less intense, if we make allowance
for the lower temperature at which these reactions take place and
which would naturally retard the reaction considerably. In Lum-
bricidae the heterotransplant may remain alive in toto, while even
in amphibia part of the heterotransplant becomes necrotic. In
adult invertebrates also heterotransplantation succeeds only in
certain, not too distant, species. We studied recently the species
differential in such simple forms as the blood cells of arthropods.
We found that the blood cells of limulus are specifically adapted
to their own blood serum. Their activities are optimal in Limu-
lus serum. Limulus serum surpasses any other serum so far
tested.

In order to determine this relationship we made use of experi-
mental amcebocyte tissue and we compared the rapidity and quan-
tity of outgrowth from this tissue in different sera. The pictures
we obtain under those conditions are very similar to those pre-
sented by outgrowing embryonic connective tissue. We find then
even in the blood cells of invertebrates a specific adaptation be-
tween the species differential of the blood cells and the supple-
mentary species differential of the body fluids.

If we pass to the lower vertebrates, we note in amphibian larvae
a condition somewhat analogous to that in invertebrates. Espe-

TRANSPLANTATION AND I INDIVIDUALITY. 167

cially the grafting of amphibian embryos, a method originated by
Born and subsequently used by Braus, Harrison and others, is
interesting in this connection; furthermore the transplantation of
skin (Lewis, Weigl), of the eye (Uhlenbath) and the experi-
ments of Spemann add valuable data. There is apparently no
difference between the results of auto- and homoiotransplanta-
tions. On the other hand a heteroreaction again exists, but is less
marked than in the case of higher veretebrates. Transplantation
into nearer related species succeeds better than into further dis-
tant species.

In adult amphibia we find the first indication of the existence
of a homoiodifferential. However, in adult fishes and amphibia
the reaction against homoiotransplantation seems to be less
marked than in adult mammals and birds, although we must con-
fess the evidence concerning the fate of homoiotransplanted
tissues in these classes appears somewhat contradictory. We have
furthermore to consider the difference in temperature at which
reactions occur in lower and higher vertebrates. This factor
might render the homoio reaction much slower in amphibia.
Despite these difficulties we may provisionally conclude that the
reaction against homoio and heterodifferentials is somewhat less
pronounced in lower than in higher vertebrates.

The interesting observations of Murphy and Rons who made
successful heterografts of tumors in the allantois of chick em-
bryos suggest that the individuality reaction is absent even in the
embryo of higher vertebrates which serve as hosts. This agrees
with an observation of Braus which seems to indicate that in am-
phibian larvae used as hosts the heteroreaction appears at a cer-
tain stage, namely, when the circulation has been established in
the transplant. Likewise if we transplant embryonic mammalian
tisue into adult hosts the homoioreaction seems to be somewhat
elss pronounced in the case of certain tissues ; the reaction, how-
ever, does exist. Embryonic mammalian tissues call forth a very
rapid heteroreaction in adult hosts (Saltykow).

To summarize, we find absence of homoioreaction in inverte-
brates and larvae of lower vertebrates. The heteroreaction exists
here, but is less pronounced. In lower adult vertebrates the

1 68 LEO LOEB.

homoioreaction as well as the heteroreaction exists, but again
both are probably less pronounced than in mammals.

There has, therefore, as far as we can judge from the experi-
ments so far recorded, taken place an ontogenetic as well as a
phylogenetic evolution in the formation of the individuality dif-
ferential. Ontogenetic and phylogenetic evolution has apparently
led to an individualization of the organisms. There is added
gradually to the reaction against the species differential the re-
action against the individuality differential.

If we inquire into the mechanism on which this gradual refine-
ment depends, several factors would have to be considered :

1. The individuality differential might as yet be absent in the
lower forms, especially in invertebrates, and appear only in ver-
tebrates.

2. The individuality differential might be present in all animal
organisms, but the supplementary substance in the body fluids
might develop only in certain ontogenetic and phylogenetic stages
and consequently the reaction on the part of the host might be
very weak.

3. The individuality differential might be present and the host
might react towards it, but somehow the tissues of lower or-
ganisms show a very low degree of sensitiveness to these re-
actions. In other words, is this individualization due to the
gradual acquisition of the individuality differential or to the de-
velopment of a reaction on the part of the host against the indi-
viduality differential or to a greater sensitiveness on the part of
the higher tissues to this reaction? Authors have so far not dif-
ferentiated between these three possibilities, and the data on hand
do therefore not permit us to decide the question definitely.

There exist, howrever, some facts which have a bearing on this
question. In the first place we find a very interesting exception
to the statement that individuality reactions are absent in lower
forms: in a certain rhizpocl. Orbitolithes, Max Schultze observed
as early as 1863 a peculiar reaction which was further analyzed
by P. Jensen. Two pseudopodia belonging to the same individual
unite, the protoplasms flowing together at the point of contact ;
but if two pseudopodia belonging to two different individuals
touch each other a contraction occurs and a reunion fails to take

TRANSPLANTATION AND INDIVIDUALITY. 169

place. An individuality differential seems to be present in these
protozoa and to lead to a marked direct interaction of proto-
plasms. There exist, perhaps, a few similar reactions in certain
other unicellular organisms. In other similar protozoa, however,
the reaction has not been observed by more recent investigators.

If it were possible to generalize from this observation, we
should conclude that the individuality differential is present in all
animal tissues ; and that the lack of reaction must be due to one
of the two other causes. On the other hand there are some indi-
cations that the chemical constitutions of the cell proteins of the
embryo differ from those of the fully developed forms. To men-
tion only one fact of this kind: while according to Roessle em-
bryonic tissue contains the same antigen for hemolysins as the
adult tissue, it lacks according to Braus the substance which after
injection into animals of a different species or class calls forth the
production of precipitins ; neither does embryonic tissue bind
precipitin which had previously been produced through hetero
injection of adult material. Probably the precipitin reaction is in
this case a less reliable indicator of the presence of the differen-
tial than the hemolysin reaction, which seems to be positive. It
is possible that the absence or presence of the supplementary sub-
stance in the body fluids may at least in part, be responsible for
the hetero reaction ; this is suggested by the observation of Braus
according to whom extremities of amphibian embryos grafted on
hosts which are as yet in an embryonic stage show the effect of
the unsuitable soil, as soon as the body fluids of the host begin to
circulate in the grafts.

If we take all these facts into consideration, it seems on the
whole more probable that the individuality differentials exist in
all animal organisms, but that the individuality reactions are lack-
ing in lower forms.

INDIVIDUALITY AND SPECIES DIFFERENTIAL AND FERTILIZATION.
There exists an interesting correlation between the lack of the
individuality differential in lower ontogenetic and phylogenetic
forms on the one hand and between the lack of an individuality
reaction on the part of egg and sperm chromosomes throughout
all classes on animals. Homoiofertilization is the normal occur-

I7O LEO LOEB.

rence and it can be considered as an intracellular transplantation.
Homoiotransplantation succeeds in this case as well as autoreac-
tion. Indeed it is a transplantation which usually occurs under
homoio conditions. Furthermore, just as in the case of the more
primitive tissues a reaction of incompatibility occurs after hetero-
transplantation, in a corresponding way a reaction of incompati-
bility occurs, when chromosomes of germ cells with different
species differentials meet, as the experiments of Baltzer, Tennent,
Moenkhaus, Guyer, Newman and others have shown. With this
conclusion is even in agreement the cross-fertilization between
widely different classes which have been made possible through
the discovery by Jacques Loeb of a method permitting such
hybridizations. In this case the incompatible chromosomes are
eliminated.

There is an additional similarity between heterotransplantation
and heterofertilization. While in both cases the incompatibility
increases on the whole in correspondence with the distance of the
species, there is no absolute agreement between the effect of
heterofertilization and heterotransplantation on the one hand and
the relationship of the species on the other hand and in both cases
reciprocal relations may lead to very divergent results.

There exists on the other hand, as far as we can judge from a
study of recorded hybridizations, one noticeable difference be-
tween the reactions of chromosomes and tissues. In the latter
there is an indication that the incompatibility between species dif-
ferentials increases in the course of ontogenetic and phylogenetic
development. A study of hybridization on the other hand does
not suggest a greater mutual tolerance of chromosomes of dif-
ferent species in invertebrates as compared to those of verte-
brates.

LACK OF THE HOMOIO SENSITIVENESS IN CERTAIN MAMMALIAN

TUMORS.

We have so far assumed that all mammalian tissues possess
the individuality differential and call forth reactions against the
homoiodifferentials. In general this statement is correct. There
exists, however, at least one notable exception to this rule.
There are certain tumors which are able to grow with the same

TRANSPLANTATION AND INDIVIDUALITY. 1 71

readiness in another individual of the same species as in the or-
ganism in which they originated. This applies by no means to all
tumors, on the contrary the large majority of tumors behave in
this respect about like ordinary normal tissues which succumb to
the homoioreaction, and are only able to grow in the individual
a part of which they formed originally. Why there is a relatively
limited number of tumors which behave differently, why they
are able to withstand the injurious influences of homoiotoxins
which destroy the large majority of tumors, we do not know defi-
nitely. It is, however, very probable that these tumors also pos-
sess the individuality differential, and that their ability to with-
stand the homoiotoxins is due to their diminished sensitiveness
combined with an increased growth energy. Very often the
" homoioreaction " which is ordinarily lacking in these tumors
can be called forth through previous immunization of the host.

It is exactly such tumors which lack the individuality reaction
which have been used by Tyzzer, myself and Fleisher and again
by Tyzzer and Little in order to study the inheritance of tol-
erance for grafted tumors. Tyzzer crossed for this purpose
white and Japanese waltzing mice; Fleisher and myself used
diverse strains of white mice. The strain differential is further
distant in the spectrum of relationships than individuality differ-
ential ; and Tyzzer and Tyzzer and Little dealt even with some-
thing akin to species differentials. Tyzzer believed that such
investigations may throw light on the character and origin of
tumors. While this view is probably not tenable, these investiga-
tions throw some light on the hereditary transmission of strain
and species differentials. I found here again an intermediate
mode of inheritance, while Tyzzer and Tyzzer and Little found a
more complex mode of inheritance. Again I suggested in inter-
preting my own as well as Tyzzer's results the presence of mul-
tiple factors, an interpretation which was subsequently adopted
by Tyzzer and Little.1

1 In a paper which just appeared — after completion of this manuscript (C.
C. Little, Journal of Experimental Zoology, 1920, XXXI., 307) — Dr. Little
expresses the opinion that the intermediate type of heredity is not the typical
mode of inheritance of the individuality differential. I believe that the views
which Dr. Little expresses are based on a lack of distinction between species
or strain differential on the one hand and individuality differential on the

172 LEO LOEB.

INDIVIDUALITY DIFFERENTIAL AND POTENTIAL IMMORTALITY OF

SOMATIC CELLS.

The fact that certain tumors can withstand the action of the
homoiotoxins has a still wider bearing. We must remember that
common transplantable tumors are the direct descendants of ordi-
nary tissue cells, such as we normally find in the individuals of
the particular species which we use. The tumors may be derived
from a variety of normal tissues and in general the transforma-
tion from normal cells into tumor cells takes place under the in-
fluence of a long continued action of various factors enhancing
growth. Tumor cells are therefore merely somatic cells which
have gained an increased growth energy and at the same time
somehow gained, in some cases, the power to escape the destruc-
tive consequences of homoiotoxins. This ability of certain tumors
to grow in other individuals of the same species has enabled us to
prove through apparently endless propagation of these tumor
cells in other individuals that ordinary somatic cells possess the
potential immortality in the same sense in which protozoa and
germ cells possess immortality. Thus tumor transplantation
made possible the establishment of a fact of great biological in-
terest which because of the homoiosensitiveness of normal tissues,
could not be shown in the latter.

We wish, however, especially to emphasize the fact that our
experiments did not merely prove the immortality of tumor cells,
but of the ordinary tissue cells as well, the large majority or all
of which can be transformed into tumor cells. At an early stage
of our investigations we drew, therefore, on the basis of these
experiments, the conclusion that ordinary tissue cells are poten-
tially immortal ; notwithstanding the fact that especially under
Weismann's influence the opposite view had been generally ac-

other. I have referred to this distinction in this paper ; I have also empha-
sized how unsuitable transplantable tumors are for the analysis of individ-
uality differentials. In my previous papers I have discussed the influence of
such adventitious factors as sex, age and pregnancy on the individuality dif-
ferentials, and showed that within the limits of our experiments such factors
do not noticeably influence the individuality reaction ; this applies to guinea
pigs above the age of four weeks, as far as the age factor is concerned. We
also referred to the apparently abnormal behavior of tissues of the child trans-
planted to the mother. The factors that are responsible for this peculiarity
need, as I stated above, further investigation.

TRANSPLANTATION AND INDIVIDUALITY.

173

^o
rt

U)

_OJ

'o

OJ

2 ii
Y S

S E

O
+j

E

rt J

"SB-

tft a)

C >>.Si
rt "C u

i rt i
i o D.

tn Hi

c tc

o
S rt
^ e

S

o rt
— i— i

JO

s

V

C
rt

CO

u

IH

C

C

I

o,

tn
C
a)

H

O. o

2 '3 w

O O n)

'o '3 t;

B 5 »

O' C

-t-J rt

I 4->

5 G

3 —

< a

w
£

o

o,
c.

E
O

IH

3
O

C
1C

•a

c

n)

a;

Q

u
nj

C.

174 LEO LOEB-

cepted, and as it seems to us, with full justification, inasmuch as
no facts were known at that time which suggested the immor-
tality of somatic cells. It was the apparently endless transplan-
tation of tumor cells which proved the contrary view.

To recapitulate what we stated above : tumors are merely trans-
formed tissue cells. All or the large majority of adult tissues are
potential tumor cells. Tumor cells have been shown experi-
mentally to be potentially immortal, therefore tissue cells are
potentially immortal.

This wider conclusion I expressed nineteen years ago. Quite
recently the immortality of certain connective tissue cells has
been demonstrated by Carrel through in vitro culture of these
cells. Under those conditions the tissue cells escape the mechan-
isms of attack to which the homoiotoxins expose the ordinary
tissue cells in other individuals of the same species. Under these
conditions the reactions of the host tissue against homoiotoxins
which would have taken place in vivo, are eliminated. We must,
however, keep in mind that this method of proving the immor-
tality of somatic cells applies only to one particular, very favor-
able kind of cells and it is very doubtful, if by cultivation in vitro
the same proof could be equally well supplied in the case of other
tissues. On the basis of tumor transplantations on the contrary
we were able to show that a considerable variety, perhaps the
large majority of all tissue cells possess potential immortality.

GROWTH CURVES AND SPECTRUM OF RELATIONSHIPS.

We may approximately represent the effect of syngenesio,
homoio, and heterotoxins on various kinds of tissues in the form
of curves, where the base lines indicate the spectrum of relation-
ships and the ordinates growth energy of the tissues in the various
hosts. We find then that embryonic and adult invertebrate
tissues and the embryonic tissues of lower vertebrates from one
class (Curve i). This, however, does not necessarily imply that
all these tisues behave in an identical manner, but that there exist
some essential similarities. Our data are as yet by no means
complete in this respect.

Very similar to this curve of the primitive tissues is that repre-
senting the growth of the transplantable tumors (Curve II).

TRANSPLANTATION AND INDIVIDUALITY.

175

CC
CU
T3

C
O to

<+-« CJ
O W

.2 °

i?

t*- o

*o

a
Q

c ^
"a 3

X 4_)

« 'I-

•^ ^
o o
n •?

9 7,

c c 2

2 -2 cd

i s'S
£§£

ig'Sg

* c

_ to

4) CD

u

o 5 -2 «

•3^3-0 2

S & c m
3 a .2 •«

s ss§

to

C

a

OZ.cn
'35-0

S c £

° -2 JS

K o. "~"

§ n
co e

C 5
2 »

V c -S

O O oj

'o '"* ""

I.
ffil

CO

u

o

.Q
cs

'c
nJ

"H,

CO

C
cS

0) to i— i

a

JJ rt C

C ^? o
tU Q- .—

g>|-
c^2

C!
US

"E, c
c o

c« 'Z3

V 5
o

K

D
O

176 LEO LOEB.

The latter, however, differ from the former in their sensitiveness
to strain differences within the same species. In addition there
may furthermore perhaps be found some differences in the be-
havior towards heterotoxins of invertebrate and embryonic tissues
on the one hand and of transplantable tumors on the other hand.
From these curves differs very markedly the curve of the adult
tissue of the higher vertebrates and similar to this is the curve of
the large majority of the tumors, namely of those which generally
are not transplantable, although in a limited number of individ-
uals of the same species they may perhaps grow (Curve III.).
The adult tissue of amphibia and fishes represents a transitional
condition between type I. or II. and III.

CELLULAR AND PSYCHICAL DISCERNMENT OF INDIVIDUALITY.

We have shown that the cells of our body are able to discern
in a quantitatively graded manner not only the difference between
their own kind, between the constituent parts of the same indi-
vidual on the one hand, and the cells of other individuals of the
same species, on the other hand, but that they are able even to
recognize in a graded manner degrees of relationship between
members of the same family. We found especially the lympho-
cytic reaction a quantitative indicator of this relationship. We
must therefore conclude that there are graded biochemical differ-
ences within the same family which these individual cells discern,
and to which they react. These reactions represent as far as we
are aware, the finest biochemical reaction known at the present
time and on the basis of these reactions we may in a tentative
manner postulate a graded system of contact substances which
regulate the interaction of various tissues.

To return in conclusion to the starting point of our discussion,
namely, the usual meaning of individuality, we saw that in the
main, it designates a social-psychical way of reaction. We are
able to differentiate between individuals as a result of certain
functions of our central nervous system. If we now inquire
how far the development of this kind of individualization is par-
allel to the power of cells of higher vertebrates generally to dis-
cern individuality, we are handicapped by the lack of data as to
the power of animals to discern not merely members of a species,

TRANSPLANTATION AND INDIVIDUALITY.

177

i

C

^ ^H •

CO *J 03 —

C w cu '"

Sf2 -1

6 g & -

aJ'-C

*j ts •«

•

CD •<-• ^

s

O

in

C ' .2

"*-^'C o

i> « " 5

2 §1 ^

"GJ ^ ^

33 ~ c

rt

S S1

JS c

c. ca

CO V-i

go c
is -s o

o 2 c

1 1

o

K1*-!
0

cu

*

° lH ^

u S c f_

o

l_^ (Tt fj^

1 1

- Brother to brothei
and mother to
child
lyngesio-transplanti

mammals and birds

a w c

~ -i-*
o. c 3

c o "o

cs '£ rt

•V B c

o

3 1

re

C

rt

"o,

in

C

CS

in

"O

o

in

w

s

D

u

I 78 LEO LOEB.

of a litter or mother and mate, but individuals of the same species
as such. This problem does not seem so far to have been con-
sidered by students of animal behavior. At least inquiries which
I made among some prominent investigators in this field, failed
to provide any definite data which might be used in this connec-
tion. From my own observations I am very much in doubt as
to the ability of such animals as the common rodents to discern
individuality in the sense in which we defined it. There seems
to be little doubt on the other hand that higher animals like dogs
and horses have such an individuality discernment, at least to a
certain extent.

It is then very probable that the mechanisms which permit the
ordinary tissue cells to discern and to react towards individuality
have developed much in advance of that mechanism of our nerv-
ous system which permits us to recognize individuals in a con-
scious manner. On the other hand after the latter faculty has
once developed, it has reached in man a very much greater de-
gree of refinement in individualization than that exhibited by the
discernment of individuality on the part of cells in general.

LITERATURE.
Baltzer, F.

'10 Arch. f. Zellforschung, V., 497.
Born, G.

'97 Arch. f. Entwickelungsmech., IV., 349.
Braus, H.

'06 Arch. f. Entwickelungsmech., XXII., 564.
Crampton, H. E.

'oo Arch. f. Entwicklungsmech., IX, 293.
Driesch, H.

'oo Arch. f. Entw.-mech., X., 411.
Dungern, Von.

'19 Munch. Med. Wochenschr., L., VII., 293-74°-
Fleisher, Moyer S., and Loeb, Leo.

'16 Journal Cancer Research, I., 331.
Goldfarb, J.

'15 Arch. f. Entw.-mech., XL.. I., 579.
Guyer, M. F.

'12 Journ. Morph., XXIII., 45.
Harms, W.

'12 Arch. f. Entw.-mech., XXXIV., 90, 1912-13, XXXV., 748.
Hektoen, L.

'07 Journ. Inf. Diseases, IV., 297.

TRANSPLANTATION AND INDIVIDUALITY. 179

Harrison, R. G.

'98 Arch. f. Entw.-mech., VII., 430.
Hedin, L. G.

'n Zeitschr. f. physiol. Chemie, L., XXIV., 242, LXXVI., 355-
Hesselberg, Cora.

'15 Journ. Exp. Med., XXI., 164.
Hesselberg, Cora, Kerwin, William, and Loeb, Leo.

"18 J. Med. Research, XXXVII f., 17.
Jensen, P.

'96 Pfliigcrs Arch., LXII., 172.
Joest, Ernst.

'97 Arch. f. Entw.-mech., V. 419.
Lewis, W. H.

'04 Am. Journ. Anat., III.
Lillie, Frank R.

'13 Journ. Exept. Zool., XIV., 515.
Little, C. C., and Tyzzer, E. e. J.

'16 Med. Research, XXXIII., 393-
Loeb, Jacques.

'12 Jour. Morph., XXIII., i.

'08 Arch. f. Entw.-mech., XXVI., 476.
Loeb, Leo.

'97 Arch. f. Entw.-mech., VI., i.

'98 VI., 297-

'07 XXIV.
Loeb, Leo, and Addison, W. H. F.

'09 Arch. f. Entw.-mech., XXII., 73.

'n XXXI., 44.
Loeb, Leo.

'15 Journ. Am. Med. Assn., L., XIV., 726.
Loeb, Leo.

'18 Journ. Med. Research XXXVIII., 393-

'18 XXXIX., 189.

'18 XXXIX., 39.

'18 XXXIX., 71.

'18 XXXII., 353-

'20 XLI., 305.

'20 Am. Naturalist, LIV., 45, 55.
Loeb, Leo.

'03 Montreal Med. Journal, July, 1903. Virchow's Archiv.

'04 CLXXVL, 10.

'10 Biochem. Zeitschrift, XXVIII., 169.
Loeb, Leo, and Moyer, S. Fleisher.

'12 Centralbl. f. Bacter., L., XVII., 135.
Meisenheimer, J.

'og-'io J. Zool. Anzeiger., XXXV., 446.
Myer, Max W.

'13 Arch. f. Entw.-mech., XXXVIII., i.

ISO LEO LOEB.

Meyns, R.

'10 Pfluger's Archiv, CXXXII., 433.
Moenkhaus, W. J.

'o3-'o4 Am. Journ. Anat., III., 29.
Morgan, T. H.

'05 BIOL. BUI.I.., VIII., 313.
Newman, H. H.

'14 Journ. Exp. Zool., XVI., 447.
Sachs, Hans.

'03 Centralbl. f. Bact., XXXIV., 686.
Sale, Llewellyn.

'13 Arch. f. Entw.-mech., XXXVII., 248.
Saltykow, S.

'oo Arch. f. Entw.-mech., IX., 329.
Seelig, M. G.

'13 Arch. f. Enlw.-mech., XXXVII., 259.
Schultz, W.

'i2-'i3 Arch. f. Entw.-mech., XXXV., 484; XXXVI., 353-

'13 XXXVII., 265. 285.

'15 XLI., 120.

'i7-'i8 XLIII., 361.
Tennent, D. H.

'12 Journ. Morph., XXIII, 17; Journ. Exp. Zool., XII., 391.
Todd, Charles, and White, R. G.

Proc. Roy. Soc., 1910, Series B, LXXXII., 416.
Tyzzer, E. E.

'15 Journ. Med. Research, XXXII., 331.
Tyzzer, E. E., and Little, C. C.

'16 Journ. Cancer Research, I., 387.
Uhlenhuth, Edward.

'13 Arch. f. Entw.-mech, XXXVI., 595.
Weigl, Rudolf.

'13 Arch. f. Entw.-mech., XXXVI.. 595.
Wetzel, G.

'95 Arch. f. Mikr. Anat., XLV., 273.
Wodsedalek, J. E.

'16 RICH.. Rn.i... XXXI.. i.

Vol. XL. April, 1921. No. 4.

BIOLOGICAL BULLETIN

AUTHORS ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, MARCH 14, 1921.

THE ECOLOGY AND LIFE-HISTORY OF AMPHIGO-
NOPTERUS AURORA AND OF OTHER VIVIPAR-
OUS PERCHES OF CALIFORNIA.

CARL L. HUBBS,
MUSEUM OF ZOOLOGY, UNIVERSITY OF MICHIGAN.

TABLE OF CONTENTS.

PAGE.

Introduction 1 8 i

Ecology 183

Breeding season 185

Sex-ratio in adults, young and embryos 186

Early differentiation of the sexes, and natal maturity of the males 187

Copulation, and the storage of spermatozoa 189

Embryonic development and natal metamorphosis 191

Period of breeding of females of different size and age 194

Number of young born by females of different size and age 195

The seasonal marks ( annuli) on the scales 197

The metamorphic annulus 200

Comparative size of the sexes at different ages 201

Rate of growth 203

Bibliography 208

INTRODUCTION.

As the life-history of these fishes is intimately correlated with
their viviparity, it may be of interest and pertinence to consider
first some of the main features of viviparity in fishes. Most
fishes are characterized by the prolific production of ova that are
fertilized in an almost fortuitous manner. Within the group,
however, some form of protection of the eggs has become re-
peatedly evolved, in correlation with a decrease in the number
of ova and with a less random fertilization. This protection is
variously accomplished by one or both parents, — by the burying
of the eggs in relatively safe situations; the construction of
nests of gravel, plants or bubbles ; the driving of predatory

181

1 82 CARL L. HUBBS.

enemies away from the eggs or young; the gestation of the young
within the mouth or blood pouch ; the enclosure of the eggs in
a tough capsule, or finally by the actual development of the
young within the oviduct or ovary of the mother.

The degree to which this viviparity has become perfected varies
widely in the different groups of viviparous fishes. Some
teleosts, such as the scorpaenoid fishes, give birth to thousands
of minute embryos, still nourished by a relatively large yolk-sac ;
while others bear only a few young, but fully developed and
capable of self-support, almost immediately after birth, in the
normal manner of adult fishes. In some of these, as the
Pceciliidae, the embryos are nourished by the yolk in the egg, and
a meroblastic type of cleavage persists. In the Embiotocidae or
viviparous perches on the other hand, the yolk is greatly reduced
in bulk, the cleavage approaches the holoblastic type (according
to Eigenmann, 1894, etc.), and the embryos are profoundly modi-
fied structurally.

The viviparous perches (see Figs. I and 2) comprise a com-
pact group, the family Embiotocidse (and the suborder Hol-
conoti) of the Acanthopterygii or spiny-rayed fishes. The group
is relatively old, apparently, for the many structural features cor-
related with viviparity are common to all of the species, and hence
became fixed before the extensive generic differentiation char-
acteristic of the family arose. Almost all of the species are
generically distinct from the others, another situation suggesting
the age of the group (cf. Eigenmann and Ulrey, 1894; Jordan
and Evermann, 1898, and Hubbs, 1918). The immediate rela-
tionships of the Embiotocidse not being apparent, nothing definite
can be said concerning the origin of their viviparity.

The viviparity of the embiotocids was first definitely made
known by Dr. A. C. Jackson in 1853, in a letter to the elder
Agassiz. These fishes then almost immediately attracted the at-
tention and study of a number of zoologists, among whom may
be mentioned both Louis and Alexander Agassiz, Gibbons, and
Girard. Later Ryder, and particularly Eigenmann, studied their
embryology, and Jordan and Gilbert, Eigenmann and Ulrey and
others also studied the group (see bibliography). I have re-

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

cently reviewed the family from a taxonomic standpoint (Hubbs,
1918), and have studied the life-history of several species, that
of Amphigonopterus aurora (Fig. 2) in greatest detail. Although
the following account is largely based on this species, comparisons
with others are made in several connections.

ECOLOGY.

Nothing whatever has been written concerning the life-history
of Amphigonopterus aurora, the only species of the genus, and all
that has been printed concerning its environmental relations is the
statement that it is an inhabitant of the tide-pools of Monterey
Bay, California, and that it feeds on algae. Its habit of feeding

FIG. i. Adult female of Micrometnis minimus.

on algae, except when very young, when I found it feeding on
copepods, is correlated with its tricuspid teeth and comparatively
elongate intestine. Similarly I found that the related Micro-
metnis minimus (Fig. i), though chiefly herbivorous, feeds on
small crustaceans when young (Pt. Loma ; December 31), and
occasionally is caught on clam bait when adult. These two species
alone comprise a distinct subfamily, the Micrometrinae, which I
have recently distinguished (Hubbs, 1918).

Amphigonopterus aurora inhabits the tidal pools and channels

1 84 CARL L. HUBBS.

along the rock-bound portions of the central California coast,
its habitat differing widely from that usual to the . species of
the family, most of which live in the surf along sandy beaches,
or in sheltered bays. In the preference of this species for this
extreme type of habitat it is perhaps most nearly approached by
Micrometrus minimus. The associational distribution of even
these two species is, however, imperfectly complementary. Both
range from the region of San Francisco southward in the cold
coastal waters of central California to the reefs about Point
Conception, where the habitat of A. aurora is abruptly terminated,
whereas that of M. miniums is continued southward in the rela-
tively warm waters along the coasts of southern and of Lower
California. Micrometrus minimus is in fact most abundant in
the warmer southern portion of its range, although fairly com-
mon northward, where the ranges of the two species coincide.
Here, however, M. minimus occupies to a large extent a biotic
association different from that of its congener, but adjacent to
it : it lives and breeds in or not far below the lowermost tide-
levels, mostly in the low, deep, plant-filled pools of the reefs (but
also in enclosed bays and estcros}. Amphigonopterus aurora, in
contrast, is restricted to the reefs, and while breeding at times
even in the same pools with its relative, more commonly lives and
breeds in the pools and channels of medium tidal height, particu-
larly those that are largely open, free of eel-grass and algse, and
floored with sand. The breeding season of Amphigonopterus,
moreover, appears to begin earlier than that of Micrometrus in
central California (see following section). Both of these fishes
were found associated in the lower outer rock-pools of the Cali-
fornia reefs with the following other species of the family : Em-
biotoca jacksoni, E. latcralis, Hypsurus caryi and Cymatogastcr
aggregatus.

In these open pools the fishes of this species swim about freely
in schools1 at low-tide, occupying the mid-water stratum chiefly,

1 That these schools of A. aurora remain intact for considerable periods of
time appears probable from two sets of observations. A number of pools on
the reefs just south of Piedras Blancas, and just south of Pt. Sal, by careful
observation over a period of several days (during a single series of low-tides
in each case), were found to contain many more individuals than any of the
adjoining pools, and to contain schools of apparently the same individuals,

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA. I 85

occasionally leaping clear of the water (a habit observed only at
Pt. Purisima, California, on August 13). During high-tide, how-
ever, they must seek the protection of crevices in the sides of the
pools, for otherwise they would be dashed about on the rocks by
the pounding, churning surf as it breaks on the reefs. In corre-
lation with its preference for pools clearer of vegetation, and
with its habits of swimming about rather more freely, Amphi-
gonopterus aurora- is more extensively silvery than Micrometrus
minimus, and lacks the dark color pattern characteristic of that
species2 (see Fig. i). In this connection we should recall that
practically all free-swimming or pelagic fishes are silvery and
lack the dark markings usually developed in fishes which live
among rocks or plants.

Like most reef-fishes examined, Amphigonopterus aurora is
not heavily parasitized, a fact apparently correlated with the
strength of the wave and tidal currents on the reefs. Occasion-
ally, however, a slender lernsean copepod was found attached to
the inner surface of the base o>f the pectoral fin, or to the anal fin
near its base.

BREEDING SEASON.

The breeding season of Amphigonopterus aurora is the sum-
mer, approximately synchronous with that of Cymatogaster ag-
gregatus, which breeds in bays and estuaries. It begins shortly
before the first of June, as is evident from the observations made
on the reefs of Piedras Blancas, California, during the first week
of that month. Most of the many females taken on that occasion
contained young, relatively few of the largest being spent. Fur-
thermore, all of the hundreds of young in the higher pools were
approximately of the size at which they are born. The breeding

judging from the approximate number of the fishes of each size. In a number
of pools fished during the summer and fall, the young of the year of each
sex were so uniform in size that it seemed probable that they had remained in
that pool together since their birth at some time during the breeding season.

2 The most conspicuous color feature of Amphigonopterus aurora (the one
on which its specific name was based) is the longitudinal band of golden or
orange color, which is rarely obsolete (a row of blotches of similar color and
position often is present in M. minimus, representing this longitudinal band
of A. aurora). In young specimens the vertical fins are dusky with a reddish
tinge, the spinous dorsal, and in the male the anterior portion of the anal fin,
being darkest ; the pectoral, nearly colorless.

1 86 CARL L. HUBBS.

season was found to be still at its height about the middle of Julyv
but to have ended long before October 26.

The breeding season of Micrometrus minimus in central Cali-
fornia apparently commences somewhat later than that of Am-
phigonoptcrus aurora. Of females taken in the same pool near
Piedras Blancas on June 2, those of Micrometrus contained
embryos from 4.6 to 18.7 mm. long, none ready for birth, whereas
those of Amphigonopterus contained embryos 12 to 34 mm. long,
the largest obviously ready for birth, being similar to numerous
young found at the same locality. A single female of Micro-
metrus collected at Point Purisima, California, on August 14,
was spent, but all of those taken during June still contained
young. Young showing considerable growth since birth were
taken in reef-pools of southern California early in July, suggest-
ing that the breeding season occurs earlier in the warmer waters
to the southward.

SEX-RATIO IN ADULTS, YOUNG AND EMBRYOS.

In the breeding pools poisoned near Piedras Blancas, the
females were found to be more numerous than the males, in the
proportion of nearly two to one: of those taken (by poison) and
sexed, 139 proved to be males; 264, females.1 An unrecorded
observation by Dr. C. H. Gilbert on Cyniatogastcr aggregates
may have some bearing on this point : he has observed a single
male " herding about " several females. On the contrary I ob-
served several small fishes of the same species, presumably males,
accompanying a mating pair (Hubbs, 1917). However this may
be, the numbers of the sexes of the young fishes were found to be
approximately equal in two pools at different localities fished in
the autumn, after the close of the breeding season : in these two
pools, 83 young males and 82 young females were obtained. The
35 adult specimens of Micrometrus minimus obtained in a single
pool near Piedras Blancas comprised 19 males and 16 females.

A more accurate sex-ratio can be obtained by determining the
sex of a series of embryos, by means of the secondary differences
in the anal fin, which become clearly evident very early in the

i This difference in the sex-ratio may be determined by the greater lon-
gevity of the females.

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

18?

development of Amphigonopterus and Micrometrus (see next
section and Fig. 2). Of 630 embryos of Amphigonopterus aurora
(from mothers one to four years old), 337 were found to be
males, 293 females (sex-ratio: 100 males to 87 females); the
ratio does not vary with the age of the parent fishes, being 100 to
86.5 in the embryos from the yearling females only. Of 150
embryos of Micrometrus minimus examined, 76 were males, 74
females ; the proportion of males to females in each of the seven
cases included was, 7 to 9, 9 to 14, 10 to 12, 10 to 13, n to 8, 13
to 12, 16 to 6. No tendency toward uniformity in sex even of
embryos lying within the same ovarian sheets was evident ; hence

Live

Sto

FIG. 2. Newly-born young male of Amphigonopterus aurora, from near Piedras
Blancas, Cal. Dissected to show mature development of testes.

polyembryony does not occur. Eigenmann recorded similar data
for Cymatogaster aggregatus, having distinguished the sexes cyto-
logically.

EARLY DIFFERENTIATION OF THE SEXES, AND THE NATAL
MATURITY OF THE MALES.

Secondary sexual differentiation is early manifest in the de-
velopment of Amphigonopterus aurora and Micrometrus mini-
mus. The differential number of anal rays characteristic of the
sexes of each of these two species (Hubbs, 1918) is clearly ap-
parent in embryos only 12 mm. long (the anal rays are first
formed when a total length of about 10 mm. has been attained).
Eigenmann was unable to distinguish the sexes in Cymatogaster

i88

CARL L. HUBBS.

cytologically at earlier stages than this, although he traced the
development of the sex-cells from much smaller embryos.

At the 15 mm. stage the sexual differences in the form of the
anal fin are also apparent in Amphigonoptenis and Micrometrus,
although the gonad remains merely a fine strand of tissue ; but at
the 20 mm. stage the testes have begun to enlarge, and a slight
thickening of the radial membranes marks the position of the

/£? x^VSNOnf

VJ&T VV\(

fc$ ' <\\(?$'£z£ A
*rf\^T,a0k 8

/ 1 • i\\\
«'.'(*•

,',V«i "

/ / . ' . *

r.irf'' 'W

"VJM^"-V'

«8) / ';.v;-:""'

\ * «

FIG. 3. Section of testis of a newly-born Amphigonoptenis, showing several

stages in spermatogenesis.

elaborate gland developed later on the anal fin of the male. The
development of these primary and secondary sexual structures
thence rapidly proceeds in Amphigonoptenis aurora until birth,
at which time the gland on the anal fin is fully elaborated (see
Fig. 2), and all stages in spermatogenesis from the primordial
germ cells to transforming spermatids at least are evident in the
testis, the spermatogonia predominating. Just after birth, the

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA. 189

transforming spermatids and spermatozoa appear most .abundant
(Fig. 3) : the testes, as well as the gland on the anal fin, are a;
well developed in these newly born males, as in the one-, two-
and four-year-old males obtained during the breeding season.
Both the primary and secondary sexual structures become greatly
reduced in size during the autumn, winter and spring of the first,
as of the succeeding years. In fact two young males only a few
millimeters longer than the birth stage, collected in June, were
already " spent." The writer has further determined that the
testis becomes similarly enlarged in Micrometrus minimus and in
Cymatogaster aggrcgatus just before birth. No evidence was
obtained, however, to indicate that the males of Embiotoca later-
alls are mature at birth ; even the two one-year-old males of this
species (in and 116 mm. long to caudal) obtained near Pt. Sal,
in California, on June 17, with breeding females, were immature.
It is quite possible that the natal maturity of the males is con-
fined to the smaller species of the Embiotocidse.

This natal maturity of the males is particularly significant in
view of the fact that the females bear young first at the age of
one year. This phenomenon, while unique in the whole class of
fishes, so far as the writer is aware, finds a physiological parallel
in protandric hermaphroditism, in the frequent early maturing
of the male (the "grilse") in the Salmonidse, and in the earlier
seasonal activities of the males of many animals.

COPULATION, AND THE STORAGE OF SPERMATOZOA.

Dr. Eigenmann (1894, p. 420) summarized one phase of his
studies of another viviparous perch with this statement : " Copu-
lation takes place in Cymatogaster during June or early July,
although the eggs are not fertilized till the following December."
He based this conclusion firstly on observations on the seasonal
activities of the two sexes, and on the seasonal development of
the testes in the male, and secondly, on the discovery of the
presence of spermatozoa in the oviduct and in the ovarian folds
of the female, during the latter part of the summer, and the
autumn. The writer has been able to extend this evidence by the
first observation of the copulation of this, in fact of any, embio-
tocid. The female of the pair in question upon capture was

IQO CARL L. HUBBS.

found to contain only one young, partially protruding from the
oviduct, and of the same size as numerous others recently born,
found swimming about in the same body of water (Hubbs, 1917).
In Amphigonopterus aurora also it is probable that, copulation
having taken place during the breeding season in the summer,
the spermatozoa are retained in the females until winter (or pos-
sibly late autumn or early spring), when fertilization occurs and
whence intramaternal development proceeds for several months ;
and that, therefore, one year elapses between the time of copula-
tion and the birth of the young. Six lines of evidence lead to, or
are at least not inconsistent with, these conclusions.

1. This condition apparently holds in C y mat og aster aggregatus,
a distantly related species in which the breeding season is ap-
proximately synchronous with that of the present species, the
structures correlated with viviparity similar, and in which the
adults, and the young at birth, are of similar size to those of
Amphigonopterus aurora.

2. Females taken in the autumn contained no young, and males
secured on October 26, November 25 and April I had small and
obviously non-functional testes, whereas all of the males taken
with the breeding females in the summer had mature testes. This
is true also of Micrometrus minimus, and perhaps of all embio-
tocids.

3. The smallest embryos, only 12 mm. long to the caudal fin,
taken from any of the females secured in June, would presum-
ably have attained their full embryonic size (about 30 to 35 mm.)
late in the summer, toward the end of the breeding season. This
fact suggests a moderately long period of gestation; the largest
females, which at this time were bearing young, presumably had
contained embryos for several months. A similar situation holds
also in the case of Micrometrus minimus.

4. The largest females, which produce young early in the sea-
son, were found to be in a spent condition, not containing a new
lot of embryos, during the months of July and August. Although
the data are less complete, this condition appears to hold also in
the case of Micrometrus.

5. The fact that the smaller yearling females bear fewer young

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA. 191

*

than do the larger ones (as also in Micrometrus miniums and
other embiotocids) suggests that fertilization is delayed for a
considerable period subsequent to copulation. Now it is a well-
known fact, in oviparous as well as in viviparous fishes, that the
larger females of a given species are more prolific than smaller
ones, and that is the situation in this family. In the present in-
stance, however, it is presumed that the copulation preceding the
development of the first brood of young (those carried by the
yearling females under discussion) takes place soon after birth,
when there is little variation in size or age (see preceding sec-
tion). If fertilization were then immediately effected, some
rather anomalous method of fertilization, or of egg production or
resorption, would have to be postulated to explain why those
females which would be smaller at the end of pregnancy bear
fewer young than those females, which for some reason, early
birth or otherwise, are destined to be larger when their young
are ready for birth. But if it be assumed that fertilization is de-
layed for some time, until a considerable variation in size shall
have arisen, the bearing of the fewer young by the smaller year-
ling fishes becomes no longer such a special problem.

6. Finally, the young males are sexually mature immediately
after birth (see preceding section), at a time when they are asso-
ciated only with the newly-born females, which apparently do not
bear young until the next breeding season, one year later. A
similar situation presumably holds in the cases of Micrometrus
minimus and Cymatogastcr aggrcgatus.

EMBRYONIC DEVELOPMENT AND NATAL METAMORPHOSIS.

The developing embryos of the Embiotocidse, in compensation
for the reduced amount of yolk in the relatively minute egg, de-
rive their nourishment almost entirely from the nutritive ovarian
fluid in which they are bathed. This fluid, as Eigenmann (1894,
etc.) determined, is circulated by the action of cilia through the
embryos. The portion of the alimentary canal in which absorp-
tion chiefly takes place is doubtless the hypertrophied hind-gut,
which in the embryos of Amphigonopterus and Micromctnts, as
of other genera of the family, is a wide but thin-walled tube
nearly filled with long, hollow, vascular villi. The respiration of

192 CARL L. HUBBS.

these embryos is seemingly largely effected over the surface of
the body and fins, especially in the highly elevated vertical fins, in
the distal dermal flaps of which the large interradial vessels form
an extensive capillary net-work (cf. Ryder, 1885, 1893). The
embryos, thus supplied with food and oxygen, pass through their
development in the ovary, lying tightly packed against the ovarian
walls and ovarian sheets, some directed forward, others back-
ward, in such a fashion, as Agassiz long ago pointed out, as
greatly to conserve space.

About the time of birth, the young of Amphigonopterus (and
of other embiotocids) undergo a notable change, which may be
termed the natal metamorphosis. The body becomes thicker, the
flesh firmer ; the vertical fins become shorter and less flexible,
the interradial vessels smaller, the dermal flaps obsolete, and the
hind gut more nearly normal in structure. The scales have al-
ready developed so far that they are widely imbricate, and the
chromatophores have been formed in large numbers, but the body
even in the largest embryos is very much paler in color than in
the newly born young, particularly the males, which are even
darker than the adults. In other species, as Enibiotoca lateralis,
and Hypsitnis caryi, the latter as described by Agassiz, a sharply
defined color pattern is developed before birth. The viviparous
perches thus lose nearly all traces of embryonic peculiarities im-
mediately before and after birth.

The young of Amphigonopterus aurora at birth vary in length
approximately from 30 to 35 mm. (the caudal fin excluded),
being about one third or one fourth as long as their mothers, as
in other species of the family. Among several hundred examined
early in June, the smallest free-swimming young was 29.0 mm.
long, the longest unborn embryo, 35.5 mm. In a given series of
embryos from one female, the variation in length is seldom more
than one or two millimeters ; thus the two sexes in Amphigonop-
terus are seen to be of at least approximately the same size at the
time of birth ; the differential rate of growth is wholly, or almost
wholly, postnatal.

A female of Enibiotoca lateralis, 257 mm. long to caudal,
caught near Piedras Blancas, California, on June 2, contained 26
young 46 to 49 mm. long, not quite ready for birth. A slightly

LIFE-HISTORY OF AMPHIGONOPTERUS AURQRA. 193

larger female, 265 mm. long, collected in the same pool, also con-
tained 26 young, but these were larger, 50 to 54 mm. long, and
similar to recently born young obtained near Piedras Blancas, Pt.
Sal and Pt. Arguello. Another female of this species, 200 mm.
long and 125 mm. deep (exclusive of fins) contained only ten
young, 55 to 58 mm. long, some having apparently been already
born. The newly born young obtained, all during June, varied
in standard length from 43 to 58 mm. A variation in size at birth
of at least 16 mm. is thus suggested. Possibly, however, a slight
decrease in actual length accompanies the metamorphosis of this
species (as in Albnla wipes and the eels).

In the case of Micrometrus minimus, the extreme lengths of
the embryos of a single female were found to differ normally from
o.o to 3.0 mm. In three cases, however, the variation was much
greater : in one lot of six embryos, the individual lengths were
5.5, 9.5, 10.8, 11.5, 14.3, and 14.7 mm.; in a second lot, five were
6.6 to 7.6 mm. long, a sixth, 2.7 mm. ; in the third case, all but
one of the foetuses were 12.0 to 13.7 mm. long, the abnormal one
being 9.0 mm. long, and provided with a strongly sigmoid verte-
bral column and a single eye, represented only by a mass of
black pigment. Occasionally, a male embryo was found to be
slightly larger or smaller than any of its fellows. If the 16 em-
bryos in one female, 7 were males 16.3 to 18.7 mm. long, while 9
were females 17.0 to 18.0 mm. long; the average as well as the
mean length for each sex was 17.5 mm. In another lot of 23
foetuses from one female, 10 were males 20.0 to 22.3 mm. long
(average length, 21.8 mm.), while 13 were females 20.4 to 23.0
mm. long (average length, 21.7 mm.).

Soon after birth the young of Amphigonopterus aurora leave
the lower pools in which they were born, only a few remaining,
probably for a very short time, in company with the breeding
adults. They make their way thence into the pools accessible
only at high tide, in such abundance that these pools, which are
usually of small size and shallow, not infrequently harbor aston-
ishingly large numbers of these young fishes. Such pools pro-
vide a large degree of seclusion from predatory enemies, as well
as the warmest available water, in which the rapid growth of the
first months may take place. This concentration and segregation

194 CARL L. HUBBS.

of the young may also be correlated with selective mating, espe-
cially in view of the natal maturity of the males.

§

PERIOD OF BREEDING OF FEMALES OF DIFFERENT SIZE AND AGE.

The following table gives the average length of the young

found in each of fifty one- and three-year-old1 females of Ainphi-

TABLE I.

SIZE OF YOUNG OF AMPHIGONOPTERUS AURORA CARRIED BY FEMALES OF DIF-
FERENT LENGTH AND AGE.

Average Length of Young Lengths of Females Carry- Age of Females (End
(in mm. to Base of ing Young of Foregoing of Given Year

Caudal) Length. since Birth).

12... .... 77 I.(i)

13 96 I. (i)

M 85 I. (i)

IS 84 I. (i).

16... ... 76,77,82,85 I. (4)

17 77,86,90,95 I. (4)

18 78,81,84,85,88,94 I- (6)

i9 . 8^87,92 I. (3)

20 88, 88 I. (2)

21. .. 92, 94,95,98,98,99 I- (6)

85,87,91 I. (3)

23 —

24 102; 122 I. (i) ; III. (i)

25 95, 103 I. (2)

26 103 I. (i)

27 1 02 I.(i)

28 95,96; 118 I.(2); III. (r)

29 112 III. (i)

30 —

31 126 III. (i)

32 121, 123, 128 III. (3)

33 122, 123, 128 III. (3)

34 129 HI. d)

gonoptcms, all of which were obtained near Piedras Blancas.
California, during the first week of June, 1916. Both adults and
embryos were measured when freshly caught, prior to their
preservation. The smallest young are not nearly developed to the
stage at which they are born : it is improbable that the smaller
females give little to young notably smaller than those of the
larger females.

1 The method of age-determination by scale examination will be discussed
later.

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA. 195

The foregoing table indicates clearly — and the evidence has
been confirmed by the writer by a study of material obtained at
other localities — that the smaller and younger females of Amphi-
gonoptcrus give birth to their young later in tJic season than do
the larger and older females. This is true likewise of Micro-
metnis miniums (Table II. ), and as Eigenmann (1894) has
demonstrated, of Cymatogaster and other genera of the family
(but Eigenmann made no definite age-determinations). This
delayed breeding of the smaller, younger females of AmpJiigonop-
terns and other Embiotocids may be an advantageous adaptation,
allowing the growth of the yearling females to be continued, as
the structure of the scales indicates it does, during the breeding
of the older females. Most of the females being one-year fishes
in Amphigonopterus at least, this added growth would seem to
admit of a material increase in the number of young produced.

NUMBER OF YOUNG BORN BY FEMALES OF DIFFERENT

SIZE AND AGE.

/
The following table III., based upon data obtained from 48

breeding females of Amphigonopterus aurora (all obtained near
Piedras Blancas during the first week of June, 1916), conclusively
shows that the smaller females bear fewer young than do the older
and larger ones —

the one-year-old females with 5-9 young being 76—94 mm. long

(average length, 83 mm.) ;
the one-year-old females with 10—15 young being 85-103 mm.

long (average length, 96 mm.) ;
the 3- or 4-year-old females with 16-30 young being 121-129 mm.

long (average length, 125 mm.).

Exceptional cases, excluded from this summary, are those of a
three-year female with but 9 small young, and a one-year female
with 19 young. Dr. Eigenmann (1894) has similarly found that
the smaller females of several other species of the Embiotocidse
bear fewer young than do the larger ones, and the data presented
in Table II. shows that this holds true in the case of Micrometrus
minimus.

196

CARL L. HUBBS.

TABLE II.

NUMBER AND SIZE OF EMBRYOS OF MICROMETRUS MINIMUS CARRIED BY FEMALES
OF GIVEN AGE AND SIZE (IN MM. TO CAUDAL FIN).

Locality (Approxi-
mate).

Date.

Mother.

Embryos.

Age.

Size.i

Number. Size.1

IVIontercy . .

Mar. 26-Apr. 2

* 4
4 4

4 4
4 4
*4
4 4
44
4 4
44

I.
I.
I.
I.
I.
I.
I.
II.
II.
II.
II.
III.?2

4 1
4 1
4 4

59
61

65
67
7i
73
74
86
92
92
92
1 06
107
no

113
116

7
7
7

10

9

IO

10
13
17
i?
19

22

23

24

25
19

3-7-4-3
S-o-5-5
5.0-6.0
6.0-6.6
4.6-5.4
7-0
7-6-8.3
8.6-10.6
9.3-10.2
11.6-13.0
11.6-12.7
16.6-17.6
16.4-17.3
18.0-21.0
18.6-21.6
21.6-23.5

* 4

* 4

..

..

..

1 4

t 4

..

..

4 t

1 t

II

t t

Avila . . .

May 25

III.?2

VI.?2

no
198

21

21.7-23.6

Piedras Blancas. . .

44 44
44 44
44 it

it 1 4
4 1 41

44 ; 4

« « 14

t < 1 1

* * 1 1
tf 4 1

»1 41
i 1 ( t
4 ( 4 t

June 2

*4

4 4
4 4
4 4
11
4 4
4 4
4 4
4 4

4 4
4 4

4 4

I.
I.
I.
I.
I.
I.
I.
I.
I.
I.
I.
I.
I.
I.
II.
III.

58
60
60
61
66
67
70
70
7i
73
74
76

79
79
85
97

3

2

4
6
8

7
6
8

9

8
8

12
10
13

IS
16

5-3-5-5
4.6

6-7
10.0-10.4

7-4-8-9
io.o-n.8

5-5-14-7
12.7-13.4
12.3-13.0
12.2-12.5
15.0
9.0-13.7
16.0-18.0

I5-7-I7-7
15.0-17.0
16.3-18.7

Pt. Sal

June 17

44
4 I

4 4
4 4

I.
I.
I.
III.?2

V.?2
VI.?2

69
72
74
96
114
129

6
6

7
17
23

22

ii.o-n.6
2.7-7.6

I3-2-I3-7
17.0-18.3
20.0-23.0

«.

1 1

**

«

Pt. Purisima

June 1 8

I.
I.
I.

64
67
76

2
6
6

IO.O

8.3-9-3
13.4-15.0

4 4

1 Preserved material measured.

2 Age uncertain, owing to the development of apparently accessory annuli.

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA. 197

TABLE III.

SIZE OF YOUNG CARRIED BY FEMALES OF DIFFERENT SIZE AND AGE.

Number of Lengths of Females Age of Females (End of

Young. Carrying Young. Given Year syice Birth).

5 87 I. (j)

6 76,82 I. (2)

7 77. 77. 78, 79, Si, 84, 85, 87, 88 I. (9)

8 77,81,90 I. (3)

9 84,85,86,88,94; 122 I. (5) ; III. (i)

10 92, 96, 103 I. (3)

ii 85,88,91,95,95 I. (5)

12 92 I. (i)

13 98, 98, 102 I- (3)

M 94, 95, 102 I. (3)

15 99, 103 I. (2)

16 129 IV. (i)

17 —

l8 122 IV.(l)

19 96; 123 I. (i) ; III. (i)

2O 122, 128 III. (2)

21

22 121, 128 III. (2)

23 123 IV. (i)

24 to 29 -

30 126 ni.(i)

SEASONAL MARKS (OR ANNULI) ON THE SCALES.

During recent years there have been conducted numerous
studies, of biological interest and economic significance, based
upon age-determinations and the computed rate of growth of
fishes. In these studies there has been developed and rather thor-
oughly tested a method of age-determination involving an in-
terpretation of the seasonal rings indicated in scales, otoliths and
certain bones ; the scales have been most widely used. It has been
demonstrated, most definitely in the salmonoid fishes, that the cir-
culi covering the surface of the scales (cf. Fig. 4) become weaker
in structure, more interrupted and more closely approximated
during each winter, apparently as a result of the lessened physio-
logical activity and retarded growth of the fish at that season. In
certain fishes which have been investigated, these winter marks
or annuli are indicated not so much by an approximation of the
circuli as by a change in their direction and an interruption in

198

CARL L. HUBBS.

their course, along a line parallel with the scale margin (cf. par-
ticularly Taylor, 1914) / The reason for the formation of this
type of annulus is the fact that the circuli toward the end of each
year's growth gradually become more strongly curved, whereas
those marking the new growth are straighten This is the type
of annulus formed on the scales of the Embiotocidse (see Fig. 4),

(1913)

FIG. 4. Lateral field of a scale from a three-year-old female of Amphigonop-

tents aurora, showing the annul i.

although an approximation of the circuli is also frequently evi-
dent, especially on the basal and lateral fields.

That the annuli on the scales of Amphigonopterus aurora are
formed during the winter is evident from a consideration of the
following facts. A series of young from New Monterey col-
lected on October 26, and another lot from near Pillar Pt.. Cali-
fornia, obtained on November 25, show no trace of an annulus
at the margin of their scales, and had not yet attained the com-
puted length at which this mark had been formed in larger speci-
mens. Excluding recently born young, the smallest examples of
either sex among those taken near Piedras Blancas during the first
week of June, and also the young specimens secured near Pillar
Pt. on April I, have a single annulus on their scales, some dis-

1 The statement by Taylor that no approximation of the circuli occurs in
the annuli is partially erroneous (particularly as it applies to salmonoid fishes),
as is also his conclusion that the annuli of the fishes which he studied were
formed during the summer (even Taylor's own data indicates the contrary).

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA. 199

tance within the margin. Thus it appears that considerable
growth had taken place since the annulus was formed. The
actual amount of this growth (determined by a method discussed
later), in 50 of the specimens from near Piedras Blancas, was
estimated to have varied from 10 to 28 mm. ; in more than half
(28) of these the growth had been 14 to 18 mm., while the
average growth computed to have occurred between the forma-
tion of the first two annuli of older fishes from the same locality,
was about 24 mm. No annulus was found on the margin of the
scales of gravid nor recently spent females, indicating that it is
not a breeding mark.

In certain other species of the family, the annuli are doubled
in a confusing fashion, suggesting the possibility that two annual
checks in growth are registered on the scales, one during the
winter and the other during the breeding season. For instance,
the scales of a 200 mm. female of Embiotoca laterdlis, taken on
June i/, when bearing young, show five typical winter annuli, and
in addition to these, and located between them, less distinct but
similarly formed rings, the outermost at the extreme margin of
the scales. Similarly in Micromctrus minimus the annuli are
often closely approximated or doubled (beyond the second winter
annulus) ; in these cases also the outermost annulus is located at
the margin of the scales of females carrying young. Such a
condition is seldom apparent in Amphigonopterus, but may have
introduced an occasional error in the interpretation of the scales
of the fishes three or four years old.

The annuli or seasonal rings on the scales of Micromctrus
minimus closely resemble those of Amphigonopterus aurora (ex-
cept in the more frequent appearance of doubled annuli, as just
noted). The outermost annulus is located at some distance
within the margin of the scales of yearling specimens taken in
late spring and early summer in central California. In several
specimens of both sexes, young of the preceding summer, taken
at Pt. Loma on December 31, the single winter annulus is on or
immediately within, in one male considerably within, the margin
of the scale. These facts indicate that the annuli of Micrometrus
are winter marks, that the first is formed in December in southern

2OO CARL L. HUBBS.

California, and that there is some variation in the time of their
formation.

METAMORPHIC ANNULUS.

The scales of even the largest embryos of Auiphigonopterus
aurora and of Micrometrus miniums are marked from focus to
border by evenly spaced, concentric striae ; those of all but the
most recently born young, on the other hand, are marked near the
margin by a zone in which the circuli are finer and more closely
approximated than on either side, and frequently angulated, their
course on the scale within this zone being slightly different from
that without. This mark, which is formed during the summer,
resembles the winter checks or annuli formed farther out on the
scales of older fishes, and perhaps quite as closely simulates the
annulus on the scales of the Pacific salmon. As a distinctive
name, the term metamorphic annul us is proposed for this mark.
It is likewise indicated on the scales of Cymatogaster aggregatus
and of other species of the family ; the time of its formation (as
indicated above soon after birth during the summer) has been
confirmed in the case of Embiotoca lateralis.

The cause leading to the formation of the natal annulus is ap-
parently a temporary retarding of growth immediately after
birth, just as the other annuli are supposed to be formed as a
consequence of the decreased nutrition and growth of the fish
during the cold season. In this connection there should be re-
called the sudden alteration of the method of feeding and respira-
tion forced upon the young of these fishes at birth. They are
then cast out into a very different medium, from which oxygen
must be absorbed mostly through the gills, rather than through
the skin and the tips of the fins. Instead of merely passing
through themselves the nutritive fluid with which they were sur-
rounded, they must now feed in the normal fashion of fishes. It
is obviously these changes in the manner of living, which stamp a
lasting mark on the scale. The metamorphic annulus significantly
is usually more sharply evident in the males than in the females
of those embiotocids known to be characterized by the natal ma-
turity of the males.

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

201

COMPARATIVE SIZE OF THE SEXES AT DIFFERENT AGES.

Scales were retained from numerous specimens of Ajnf>hi-
gonoptcru*; aurora of measured length, all obtained near Piedra?
Blancas, California, on June 2 and 4, 1916. Excepting a few
selected to represent extreme sizes, these were chosen at random
from the large number collected. The annuli, discussed above,
are well developed on these scales, their number indicating the
approximate age of the fish (the time of intramaternal develop-
ment, about one half year, being arbitrarily excluded). Thus the
presence of a single annulus, in addition of course to the natal
annulus, indicates that the fish was born in the preceding summer,
and that it is just completing or has just completed, the first year

TABLE IV.

LENGTH TO CAUDAL BASE OF FEMALES OF AMPHIGONOPTERUS AURORA OF

DIFFERENT AGES.

Age,

One Year.

Age, Three

Years.

Age, Four Years

Age, Five Years.

Age, Six Years.

76

mm. (i)

116 mm

(I)

122 mm. (i) 141 mm. (i) 138 mm. (i)

77

(3)

117

123 (i)

78

. . .

118

(2)

129 (2)

79

(i)

119

(I)

136 (i)

80

. . .

I2O

(I)

81

(3)

121

(I)

82

(i)

122

(2)

83

123

(I)

84

(3)

124

85

(4)

125

(3)

86

(i)

126

(I)

87

(2)

127

88

(3)

128

(2)

89

(i)

90

(2)

9i

(I)

92

(3)

93

(I)

94

(2)

95

(4)

96

(2)

97

(I)

98

(I)

99

(2)

IOO

(2)

101

IO2

(2)

103

(2)

IO4

105

(I)

I O6

107

108

(i)

2O2 CARL L. HUBBS.

of its free-swimming life. As these fishes were obtained in the
summer of 1916, those with three annuli were born in the summer
of 1913, etc.

The collecting done near Piedras Blancas and at other locali-
ties indicated that the males average decidedly smaller than the
females. It is of interest to have the field observations definitely
confirmed by age-determinations.

»

TABLE V.

LENGTH OF MALES OF DIFFERENT AGES.

Two Years. Four Years.

82 mm. (i) 89 mm. (i)

One

Year.

59 mm. (i)

60

(0

61

(0

62

(0

63

(i)

64

(3)

65

(4)

67

68

(4)

69

U)

70

(2)

7i

(i)

72

(3)

73

(0

The wide difference in the size of the two sexes of Micrometrus
minimus as well as of Amphigonopterus aurora (Table VI.) ap-
pears to be of particular significance in the case of a small vivi-
parous fish. The female of Amphigonopterus carries from 5 to
30 young which attain before birth more than one fourth the
length of their mother, whereas the testes of the male are rela-
tively small for a fish, — a fact determined by the conservation of
spermatozoa, correlated with copulation (similar size relations
prevail in certain other and probably in all embiotocids, and in
many of the viviparous poeciliids). The differential rate of
growth producing the relatively smaller size of the adult males is
entirely or almost entirely postnatal, as previously indicated ; in
this connection it should again be recalled that at birth, only the
males are mature.

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

203

TABLE VI.

THE COMPARATIVE SIZE OF THE SEXES IN AMPHIGONOPTERUS AURORA AND

MlCROMETRUS MINIMUS.

Species.

Approximate Locality
(California).

Date of
Collection.

Age
(Win-
ters).

Length in
Mm. to
Caudal.

Sex.

Number of
Specimens.

A. aurora. . . .

Piedras Blancas

June 2-4

I.

59-73

cT

25

* *

I.

76-108

9

50

* *

* t

II.

82

0"

I

'*

"

It

III.

II6-I28

9

15

' *

1 1

IV.

89

<?

I

IV.

122-136

9

5

A. aurora

Pacific Grove

II

I 1A— I7O

9

2

A . aurora ....

Pillar Pt.

Nov. 26

o

48-70

<?

18

"

o

41-56

9

17

M. minimus .

Pt. Loma

Dec. 31

I.

51-52

c?

2

I.

63-64

9

2

M. minimus .

Piedras Blancas

June 2

I.

52-62

cf

15

*

11

I.

58-79

9

14

i

* '

II.

66

cf

I

i

"

II.

85

9

I

i

"

III.

68-71

0*

3

< <

III.

97

9

i

M. minimus .

Pt. Sal

June 17

I.

62

cf

i

•

t t

I.

69-74

9

3

RATE OF GROWTH.

The rate of growth of a fish which develops seasonal marks
on the scales can be readily computed with considerable accuracy
from the known length of the fish at the time of capture and
the measurements of one of its scales. The number of scales
remaining constant throughout life, the scale increases in length
along its horizontal axis in direct ratio to the increase in the
length of the fish. Thus for a fish of given total length, the
length attained at each winter, and consequently the growth incre-
ment of each successive year, may be computed by the use of the
following formula :

length of fisli at the time of capture

length of fish at winter x

length of scale to margin

length of scale to anniiliis formed in winter x

2O4 CARL L. HUBBS.

Several possible sources of error in the use of this formula are
apparent. The scales do not develop until considerable growth has
occurred. As the length of the scale is usually measured, for the
purposes of growth computations, only from the focus cephalad
to the basal margin, an unequal growth of the different fields of
the scales mav introduce an error. There is some variation in

j

the time of formation of the annuli. The scales often overlap
less widely in young fishes than in adults, rendering the computa-
tion of size for the first winter too small. Again, the length of
the fish is measured, for growth computations, from the tip of the
snout to the base of the caudal fin. There is thus included the
length of the head, which becomes relatively shorter with in-
creased size in most fishes. As the formula assumes that the
head and body increase in the same ratio, the computed length for
the young fish at a given stage would be accurate only if the head
were then of the same proportional length. The head being rela-
tively longer, however, the computation is too small. It is doubt-
less these factors (probably all of them) which have rendered the
computed length of the young of Amphigonoptcrns at the time
of the formation of the natal annulus (just after birth) to be
less than the observed length at that stage (20 to 32 mm., instead
of 29.0 to 35.5 mm.).

TABLE VII.

COMPUTED RATE OF GROWTH OF FEMALES OF AMPHIGONOPTERUS AURORA.!

Period of Growth. Growth. Specimens

To formation of natal annulus 20—32 mm. 70

Thence to end of first winter 29-68 mm. 69

Between first and second winters 14-35 mm. 20

Between second and third winters 12-27 mm. 20

Between third and fourth winters 13—18 mm. 5

Between fourth and fifth winters 13 mm. i

Between fifth and sixth winters 4 mm. i

1 Based upon material collected near Piedras Blancas during the first week
of June, 1916.

It is evident from these figures, as well as from supplementary
data obtained from collections made on other occasions, that the
growth of the first half year (between birth and the first winter)
is greater than that of any subsequent whole year. This estimate

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

205

o
U

o
o,

CO

a

t/3

rJ

1 in

rt
in

•0 £

— U
OJ

3 S

a £
«i e

o t:
.*> 55

o
u

0

s

rt

(LI
>

£

~E

IH

12

in

in

U

0

u

J

M-,

"c

H bfl u.
bo c o

OJ «*i

be
efl

in

^3

.2 c

C91

OH 021 OOT 08 09 Ot> OZ

*( (t

tt ft

tl t(

2O6 CARL L. HUBBS.

appears particularly significant, in view of the fact that the
females become pregnant at or immediately after their first winter.
As usual in fishes, their is no evidence that the growth ever
wholly ceases during life, although it is markedly and increasingly
retarded with age.

The growth of the single four-year-old male of Amphigonop-
terus obtained near Piedras Blancas was computed from the
direct ratio of scale length to fish length. The length of the head
and body to the end of the formation of each annulus was thus
estimated to have been as indicated below.

Length at end of formation of natal annulus 26 mm.

first winter annulus 50 mm.

" second winter annulus 63 mm.

third winter annulus 73 mm.

" " " " " fourth winter annulus 84 mm.

" " " fourth year (on June 4, 1916) 89 mm.

The growth of this male, although slower, was quite similar
to that of the females, being greater between birth in the summer
and the first winter, than during any subsequent whole year; the
check in growth rate in this case follows the second, rather than
as usual the first, period of breeding.

The writer has found but one published record of a direct
observation on the rate of growth of a viviparous perch. It was
made on aquarium fishes by the late Charles Frederick Holder,
and published anonymously and without identification of species.1
The obscure but pertinent passage is as follows. " The young,
ten or twenty in number, born in the summer, are from an inch
and a half long at birtri, and attain half their adult size the first
winter, and their full growth in about two and a half or three
years." Dr. Eigenmann (1894) has remarked on the large size
of the smaller breeding females of Cymatogastcr, which he cor-
rectly assumed to be one year old. A similar rate of growth
holds in the case of Micrometrus minimus.

In the course of his extensive investigations of the life-history
of the sockeye salmon (Oncorhynchus ncrka) , Dr. C. H. Gilbert

i Another note by the same author makes it evident that the species ob-
served was Cymatogaster aggregatus (see Bull. U. S. Bur. Fish., Vol. 28, 1908
(1910), p. 1139).

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

207

(1914, pp. 61-71) has induced an important generalization, the
lazv of groivth compensation. Those salmon which grow most
in their first year (as a result of earliest hatching or of other
causes), tend on the average to grow least in their succeeding
years, while those which have attained a relatively small size at
the end of the first year, grow with accelerated speed during the
next years. The physiological mechanism of the salmon appears
to regulate its growth in such a fashion that the length of the
adult fishes of each race varies but little. It was hoped that it
might be determined whether this law of growth applies to Am-
phigonoptcrus, but so few fishes more than one year of age were
examined, that the data are incomplete. The evidence being
suggestive, however, is presented in the following table (based
upon the material from Piedras Blancas).

TABLE VIII.

COMPUTED LENGTHS OF TWENTY 3- TO 6-YEAR-OLD FEMALES OF AMPHIGOXOP-

TERUS AT THE END OF FlRST THREE WINTERS.

»

Length at End of
First Winter.

Length at End of
Second Winter.

Length at End of
Third Winter.

52

76

97

58

91

108

61

94

109

62

93

107

62

96

121

63

98

no

64

80

IOO

6?

101

US

67

94

113

68

82

109

72

93

113

74

IOI

120

75

103

119

77

IOI

I2O

79

98

124

80

1 08

120

81

96

118

83

103

I2O

85

99

III

93

109

124

Variation

52 to 93

76 to 109

97 to 124

Range

41

33

27

Range

• 57

.40

.26

Mean

J 1

It appears probable that the variation in size of Amphigonop-
tcms at the end of the third winter is less than at the end of the

208

CARL L. HUBBS.

first winter ; that the principle of growth compensation is appli-
cable to this embiotocid. Similar data gathered from one-year-
old females strengthen this conclusion.

TABLE IX.

COMPUTED SPRING GROWTH OF SPECIMENS WHICH HAD ATTAINED EITHER A
SMALL OR A LARGE SIZE AT THEIR FIRST WINTER.

Length at End of Formation of
First Annulus.

Spring Growth (in Millimeters).

10

ii

12

J3

*4

IS

16

T7

18

*9

20

21

22

23

24

25

26

27

28

Less than 75 mm. .

T

;

:

7,

3

T

I

6

2

T

5

T.

3
?.

2

2

T

4

2

J

I

T

-

-

-

I

More than 7=; mm..

BIBLIOGRAPHY.

Note: Eigenmann (1894) has given an extensive bibliography and a discus-
sion of the earlier contributions pertinent to the viviparity and embryology of
the Embiotocidas. Only the later references need therefore be listed here.
Additional papers by the same authors may be located with the aid of Dean's
Bibliography of Fishes.

Eigenmann, Carl H.

'94 On the Viviparous Fishes of the Pacific Coast of North America. Bull.

U S. Fish Comm., Vol. 12, 1892, pp. 381-478; pis. 92-118.
'95 Development of Sexual Organs in Cymatogaster. Proc. Indiana Acad.

Sci., 1894, p. 138.
'96 The history of the Sex-cells from the Time of Segregation to Sexual

Differentiation in Cymatogaster. Trans. Amer. Micros. Soc., Vol. i".

pp. 172-173-
'g6a Sex-differentiation in the Viviparous Teleost Cymatogaster. Arch.

Entwick.-Mech. Organ., Vol. 4, 1896, pp. 125-179; 6 pis.

'g6b The Bearing of the Origin and Differentiation of the Sex Cells of
Cymatogaster on the Idea of the Continuity of the Germ Plasm. Amer.
Nat., Vol. 30, pp. 265-271.
Eigenmann, Carl H., and Ulrey, Albert B.

'94 A Review of the Embiotocidse. Bull. U. S. Fish Comm., Vol. 12, 1892,

pp. 382-400.
Gilbert, C. H.

'14 Contributions to the Life-history of the Sockeye Salmon. (No. I.)

Rept. Comm. Fish. Brit. Col., 1913, pp. S3-/8, 9 plates.
Holder, Charles Frederick.

No date [Avalon] Aquarium Guide. Animals of the Submarine Gardens

Santa Catalina (no pag.)-.
Hubbs, Carl L.

'17 The Breeding Habits of Cymatogaster aggregatus. Copeia, No. 47, PP-
72-74.

LIFE-HISTORY OF AMPHIGONOPTERUS AURORA.

'18 A Review of the Viviparous Perches. Proc. Biol. Soc. .Wash., Vol. 31,

pp. 9-i3-
Jordan, David Starr.

'05 Guide to the Study of Fishes, 2 vols. [viviparity in fishes: Vol. i, pp.

124-130, Figs. 91-96. — Embiotocidse : Vol. 2, pp. 372-379, Figs. 306-

312].
Jordan, David Starr, and Evermann, Barton W.

'98 The Fishes of North and Middle America. Bull. U. S. Nat. Mus., No.

47, pt. 2 [Embiotocidse, pp. 1493-1511].
'oo Ibidem, pt. 4, plates [Embiotocidre : Figs. 577-S86].
Taylor, Harden F.

'16 The Structure and Growth of the Scales of the Squeteague and the

Pigfish as Indicative of Life-history. Bull. U. S. Bur. Fish., 34, 1914.

pp. 287-330, Figs, (bibliography).

AUTHOR S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, MARCH 14, IQ2I.

STUDIES OF THE BIOLOGY OF FRESHWATER

MUSSELS.1

EXPERIMENTAL STUDIES OF THE FOOD RELATIONS OF CERTAIN

UNIONIDJE.

WILLIAM RAY ALLEN.

CONTENTS.

PAGE.

1. Introduction 210

2. Constancy of the feeding activity 211

(a) The feeding posture 212

(fe) Hunger and the degree of digestion 213

3. The utilization of nannoplankton and megaloplankton 215

(a) Experiments in selective feeding , 216

(b) The feeding of sewage 220

(c) The feeding of infusions 221

4. The physiology of the crystalline style 223

(a) Recent studies • 223

(b) The feeding of specific substances 226

(c) Forced feeding 229

(d) The effect of temperature on style renewal • 229

5. The mechanism of ingestion 232

(a) The role of the labial palps 234

(fr) The gills as an assorting mechanism 235

(c) The marly incrustation of the shell 238

6. Summary and conclusions 238

i. INTRODUCTION.

The writer, while a member of the Indiana University Biolog-
ical Station at Winona Lake, Indiana, devoted several summers
to ecological studies of the freshwater mussels. The work was
done under the direction of Dr. Will Scott, and formed a part
of his general limnological program. A first paper on the feed-
ing habits of the Unionidse appeared in 1914. The present paper
has expanded the former through the use of experimental

1 Contribution from the Zoological Laboratory of Indiana University No.
174; submitted as a partial fulfillment of the requirements for the degree
Doctor of Philosophy.

2IO

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 211

methods, and approaches the study of the crystalline style from
the physiological standpoint. In subsequent papers I hope to
present further studies along the lines of reactions to stimuli and
distribution.

For the study of the smaller glacial lakes and their inhabitants
the region is especially favorable. Winona Lake has a maximum
length of two miles, and being of the kettle-hole type, is deep.
Its greatest depth is eighty feet. The bivalve population is
limited to a shelf about its margin. Three small creeks drain into
the lake. Their course follows the flat, marshy land between
moraines, and their volume is small and comparatively constant.
The Unionids have not extended up into the creeks. The lake
forms the upper limit of their distribution in the Winona drain-
age system.

Eight species of mussels have been recorded from this lake :
Lampsilis lutcolus, Anodonta grandis, A. cdentula, Quadrula
rubiginosa, and Lampsilis subrostratus are common, the first two
being very abundant. Lampsilis glans, Micromya fabalis, and
Margaritana marginata are rare. I have found a single specimen
of a ninth species, Quadrula undulata. All these mussels are of
the " lake " type, as contrasted with " river " mussels.

For most of the following experimental studies Lampsilis luteo-
lus and Anodonta grandis have been employed. For work during
the winter, and for comparison with the above mentioned lacus-
trine forms, mussels from White River were used. These in-
clude Quadrula hcros, 0. pustulosa, Lampsilis anodontoides, L.
ligamentimis, etc. They were collected from the east fork of
White River, near Shoals, Indiana, and from the west fork of
the same river near, Gosport, Indiana.

2. CONSTANCY OF THE FEEDING ACTIVITY.

In a previous paper I stated that the lake mussel continues
feeding at nearly all times, when under normal conditions (Allen,
'14). Virtually all the observations made since the publication
of that paper have been of a confirmatory nature. Rarely is a
freshly collected mussel found to be without food material in its
alimentary tract, often much of it wholly undigested. When

212 WILLIAM RAY ALLEN.

brought into captivity defecation usually continues to occur even
after several hours, and indicates that feeding has at no time been
long suspended. The normal position of the palps, gills, and
mantle with respect to each other is favorable to ingestion. The
structure of the mussel is such that it requires a greater effort
to refrain from feeding than to continue feeding. The ciliary
apparatus is in constant activity, regardless of the presence or
absence of food. The gills are at all times siphoning water.
Particles suspended in the water are at all times being sifted out
and caught upon the mucous secretions of the gills. The gills
have no way of rejecting such collections. They are always
passed on by the fixed ciliary tracts of the gills to a given point
on the lower margin of the inner gill which hangs between the
labial palps. Without an adverse stimulus and an avoiding move-
ment of the palps the collections pass on to their apposed surfaces.
Thence it is but a short distance to the mouth.

If, despite the structure of the mechanism of ingestion, one
might still doubt that feeding is a constant process, he is forced to
grant that feeding is speedily resumed by mussels artificially
starved. Those starved a sufficient length of time to get rid of
the crystalline style, when placed again in lake or stream, begin
the renewal of the style within fifteen to thirty minutes.

(a) The Feeding Posture.

To what extent feeding is conditioned by the position of the
mantle, palps, and gills the following experiment will show :

Some LainpsUls and Anodonta were starved several days to
insure the disappearance of the crystalline styles. The renewal
of these organs was then taken as an index of the feeding activi-
ties. Various individuals were placed in the lake, in lake water,
and in Pocahontas creek. Some were partially inserted in sand
in the normal posture, others with the siphons turned upward,
others laid upon the right or left side, and still others upon the
hinge and having the gape of the shell uppermost.

At intervals some from each situation were examined. The
styles were found to be reformed in all of them at about the same
rate. The position of the ciliated parts, with relation to one

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 213

another and to gravity, was obviously of no consequence. Food
could be passed from the gills to the palps as readily upward as
downward or horizontally. Due to the buoyancy of the mucus in
which food material is gathered and passed to the stomach, so
near the specific gravity of the surrounding water, cilia need
exert only a very slight traction upon the food masses to move
them in any direction.

No Unionids have acquired an asymmetry like that of the
oyster, but some species, particularly Quadrulse, are always found
lying upon one side or the other. They feed as readily upon
one side as upon the other. Neither the right side nor the left is
favored. There is no exceptional arrangement of ciliated parts
within the mantle chamber for countering the unsymmetrical pull
of gravity, unless the unusual size of the labial palps of some
species may be interpreted as a means of preventing the loss of
food between gills and palps.

No matter, then, in what position the gills and palps may be.
food is readily passed mouthward, and the process of food collect-
ing goes on constantly in the absence of adverse stimuli. Only
with the closure of the siphons is the streaming of fresh material
interrupted. Only in case of powerful stimulus are the palps
caused to move out of line, away from the inner gill, thus refus-
ing food masses entirely. It has been difficult to demonstrate
the constancy of the food stream upon the contiguous faces of
the labial palps. The ciliated furrows (p. 234) ha've the func-
tions of accepting or rejecting material, but to what extent the
latter function is exercised under normal circumstances it is
difficult to determine directly. When fed in high concentrations
certain specific substances were rejected entirely, others taken
readily, e.g., starch grains were never recovered in the alimentary
tract, while Glococapsa passed through in great numbers (p. 227 ).
It is permissible, therefore, to postulate that the labial palps
behave in harmony with the rest of the food-gathering apparatus.

(b) The Degree of Digestion a Response to Tissue Demands.

The cilia of the alimentary tract are virtually in constant move-
ment. (Nelson, '18, has reported the partial suspension of the

214 WILLIAM RAY ALLEN.

cilia of the style sac.) It involves little, if any, additional energy
to keep a stream of water coursing through it. That much food
material passes through without being digested is shown both by
the mass color of the feces and by the appearance of diatom and
other algal cells from the feces, when examined microscopically.
It has been argued that diatoms are accidental and not the real
food. Empty diatom tests are found in considerable quantity in
fecal matter, and there is no doubt that the contents have been
digested out of, at least, some of them. • Furthermore, the freshly
formed style, in a mussel which has been starved and then fed
with diatomaceous matter, has the amber color of diatomin. In
this case digestion is prompt and rapid. The style tends to be-
come colorless as soon as the streaming of food through the style
sac is checked by the growing style.

Since the ciliary mechanism is functioning constantly, and
since some material is digested and some is not, it may be con-
cluded that the demand for food on the part of the tissues affects
rather the secretion of digestive fluids and ferments than the con-
trol of ingestion.

On several grounds it is seen, then, that a given particle may or
may not be digested on its passage through the alimentary tract.
Nelson's suggestion ('18), that the style sac serves as a means of
returning undigested particles to the stomach where they may be
exposed again to the digestive secretions, is very plausible. The
morphology of the structure fits such a function to a nicety.
Moreover, while a starved mussel is renewing its style much green
material is being threaded into its core. The fact that some
diatoms are digested while others are not might at first thought
find a sufficient explanation in Nelson's view. For some particles
would be diverted into the style sac and others continue on un-
digested. This would perhaps as readily explain the presence of
both normal diatoms and empty tests in the rectum as my sug-
gestion of uninterrupted feeding. There is no doubt that
Nelson's explanation is entirely adequate for Ostrcca and
Modiolus, for these forms interrupt siphoning and feeding with
every tide. Their styles are absorbed and renewed regularly
with the ebb and flood. Hence twice each day the stvle sac is

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 215

thrown open to the passage of food; and for this reason, if for
no other, one should find a mingling of digested with undigested
food in the rectum. In the Unionidse this mingling does not have
an adequate explanation in the style sac. For there is no loss of
crystalline style except with starvation. No normal mussel is
found without it. Under normal conditions the style has no
color, with a very slight core of green. The amount of food
which passes through the style or style sac under ordinary condi-
tions is very slight. Hence most of the digested diatoms whose
tests are found in the rectum have been through the stomach
but once.

Resume. — In the rectum and feces the simultaneous occur-
rence of green diatoms and empty tests shows that part, but not
all, of such material is digested. At any rate it is usable as food.
Ingestion is continuous, but digestion is discontinuous, and de-
pendent upon demand for nutrition on the part of the tissues. In
the Unionidse the return of material from the intestine to the
stomach through the style sac does not account for much of the
food actually digested, for such a transference of food is pos-
sible only when the style sac is empty. Since digestion is a chem-
ical reaction the contents of the stomach should be affected alike,
and not some wholly digested and some not at all, when equally
digestible.

3. THE UTILIZATION OF NANNOPLANKTON AND
MEGALOPLANKTON.

The materials recovered from the rectum grade down from
the largest phytoplankton to very small species. While stomach
and rectal contents vary, the larger particles are obviously oftener
found undigested than the smaller. This suggests that the
smaller are more easily digested ; that everything else being equal,
digestibility is inversely proportional to size. The question
arises : does the nannoplankton constitute an important, though
less conspicuous element of the food of a mussel?

Juday (see Ward and Whipple, '18) has shown that the nanno-
plankton content of a lake may exist in vastly greater number
and in a much greater volume than does the net plankton. Even
though most of such organisms should escape through the gills

2l6 WILLIAM RAY ALLEN.

of the mussel, the recovery of only a fraction of them is sufficient
in some lakes to equal or surpass the total volume of larger or-
ganisms utilized.

(a) Experiments in Selective Feeding.

A few experiments were made to separate the nannoplankton
from the grosser, in order to feed them separately.

There is no hard and fast line by which the nannoplankton may
be distinguished, except Lohmann's ('n) arbitrary size limit of
25 microns. Stomach and intestinal contents were examined for
evident of a predominance of either larger or smaller forms.
So far as the diatoms are concerned such evidence is not very
conclusive, though on the whole the smaller organisms appar-
ently have the better of the argument. As for Flagellata, they
were commonly seen in the rectal contents, but more often in the
stomach. This is as we should expect, for some of them do not
possess so resistant a test as the diatoms.

The fact that even a few of the smaller flagellates and dia-
toms are found demonstrates that the gill-meshes are fine enough
to accomplish something with the nannoplankton, while the di-
minished number of flagellates in the rectum implies the more
complete digestion of that group. Certain flagellates are more
resistant to digestion than others, e.g., Peridiniwn; some are
doubtless, of more frequent occurrence in the alimentary tract on
account of their colonial form and greater bulk, e.g., Pandorina.

We have no very efficient mechanical means for separating
plankton according to size. The nets of finest bolting silk allow
nearly all flagellates and all but the large or colonial diatoms to
pass through. It was found that water, poured through a net
suspended in the air, forces more large organisms through the
meshes than in the case when the net is suspended in water, and
the water containing plankton poured through slowly.

A concentration of net plankton actually took place in the
plankton bucket. A complete separation of coarse from fine
plankton is not claimed, nor was it necessary to the experiment.
Undoubtedly some of the grosser forms passed through the silk.
But these were never in sufficient quantity to be detected in the

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 217

intestines of mussels feeding in the escaping water. One com-
plete experiment follows :

Water from the creek or lake was first poured through a cop-
per gauze in order to remove such debris as plant fragments and
flakes of limy incrustation. From these it passed into a conical
plankton-net of No. 20 bolting-silk, terminating in a detachable
Birge bucket. The lower portion of the cone and the bucket
were suspended in a jar of water. Measured quantities of the
natural water, varying from 50 to 250 liters in the several ex-
periments, were passed through. From time to time the meshes
of the bucket became choked and the process was slowed down.
This was especially true in the lake following high winds, when
there was considerable turbidity. At intervals, therefore, the
bucket was removed, rinsed out into a container, and the gross
filtrate brought into contact with a group of mussels which had
been starved for a few days in order to cause the disappearance
of the crystalline style. The overflow water which had passed
through the silk was allowed to siphon over into another jar in
which were kept another group of starved mussels. They were
thus given opportunity to feed upon nannoplankton almost solely.

In order to demonstrate that any regeneration of the crystal-
line style that might occur could not be ascribed to the chemical
or physical character of the water itself, a quantity of the strained
water was also siphoned over into a sheet of filter paper and mus-
sels placed in the jar beneath. The mussels fed upon filtered
water in no instance showed the least evidence of re-forming the
style. That many organisms passed through the plankton net is
well shown by the amber-green coating which soon formed on
the filter paper. Upon these organisms the second group of mus-
sels had opportunity to feed, provided the gills might be a suffi-
ciently effective mechanism to entrap particles which had passed
through the silk.

During the progress of the experiment, usually about four
hours, checks were kept in the lake or creek, near the experi-
ment. They also had been starved for the same length of time, in
tap water.

The water was taken from a depth of two to three feet in the
lake, along the east shore, where it is open to wind and wave

2l8 WILLIAM RAY ALLEN.

t

action, but has the shelter of a broad zone of Potamogeton. The
creek water was from the mouth of Pocahontas creek, having a
depth from a few inches up to more than a foot, and a bottom of
gravel and rubble. A drain from a large septic tank enters 500
feet upstream. In its upper course it receives the water of
drainage ditches in swamp land, and has a small, though constant,
volume throughout summer. Its flow is rapid, and the character
of the bottom such that mussels have not ascended it. The water
has a slightly brown color due to its swampy course (Rice, '16).

In this experiment two mussels were used as checks, three
kept in filtered water, ten placed in the overflow from the plank-
ton net, and ten were fed upon the gross concentrate given them
in tap water.

Of the two checks, one developed a well formed crystalline
style, the other none, but it had a large amount of green ma-
terial in the intestine, showing that it also had resumed feeding.

Of the ten fed upon net plankton, seven renewed the style
more or less completely, three not at all; of the naiiiioplaiikton-
fed eight had formed a style and two had not. These data are
not full, yet they demonstrate the ability of the respective food
materials to renew the style.

It will be seen that there is little difference in the power to
renew the style between the larger plankton which remained, and
the smaller which went through the silk. It is certain that the
larger forms do have a food value, for there were few of the
smaller which remained in the plankton bucket, and the residue
of grosser material alone \vas put into the aquarium with the
first group of mussels. It is possibly not quite so well demon-
strated that the other eight renewed the styles in response to
the ingestion of nannoplankton solely, but such is a reasonable
supposition. Two things at least cannot be gainsaid : ( i ) there
was a significant increase in the ratio of megaloplankton to nan-
noplankton within the bucket, and of course a reversal of this
ratio outside, without greatly affecting the crystalline style re-
newal in either case; (2) no net plankton \vas recognized in the
intestines of the second group of experimental animals, although
conspicuous in those used as a check.

The above facts do not argue a greater digestibility of the

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 219

smaller organisms. But when the matter of dilution or concen-
tration is taken into account, we have such an argument. The
greatly concentrated net-plankton was fed in a container of water
which was changed but slightly throughout the experiment. The
nannoplankton, on the other hand, was caused to stream over the
second group of mussels in the original water, and with the orig-
inal concentration. If it had been concentrated as much as the
net plankton a much more marked ingestion would probably have
been obtained. It is hoped that in the future a method for con-
centrating nannoplankton in quantity may be applied to the solu-
tion of this question.

An experiment similar to the foregoing consisted in placing
starved mussels in the creek, enclosed in a tight metal container,
whose two ends were then closed with bolting-silk. The creek
was dammed on either side of the container, so as to raise the
level slightly and maintain a flow of water through it upon the
mussels. The stopping of the meshes of the silk was expected to
interfere with the flow after a few hours. But, as a matter of
fact, there was sufficient eddy and overflow to keep the silk fairly
well washed, so that at the end of the experiment a slight current
might still be detected.

Experimental mussels and checks were placed both above and
below the sewer outlet, where they were allowed to feed for
twenty-four hours. Small numbers were used here — four mus-
sels in each situation. So far as the results derived are trust-
worthy, they show a utilization of nannoplankton as food, and
corroborate those of the previous experiment.

The mussels from the three situations showed a well-defined
gradation in the reconstruction of the crystalline style: (i) those
from experimental conditions showed only a partial renewal; (2)
the checks nearby had virtually completed the renewal; (3)
below the sewer outlet the checks had well formed and entirely
hyaline crystalline styles. Presumably they had an additional
source of food — the sewer.

The mussels used as a check, and whose styles had grown
large and hyaline, accumulated considerable masses of green in
the intestine during the twenty-four hours. Examination showed
this material to comprise, among other forms, Namcula, Oscil-

22O WILLIAM RAY ALLEN.

latoria (small amount), Sccncdesnnis, Synedra, and Tabellaria.
Considerable debris, both organic and inorganic, was found, and
some fragments of considerable length.

In the case of those fed upon " bolted " water considerable
amounts of green were found in the intestine, and the styles were
flaccid and green. No large particles were obtained. Tabellaria,
Diatoma, and Navicula occurred, but it was only very small spe-
cies in each case, and no coherent members of colonies. Very
little inorganic stuff was found, but many organic fragments,
some of them partially digested. The smaller flagellates were
proportionately more numerous than in the checks. Naturally
the diatoms were of the solider, creek type, and none of the grace-
ful lake forms adapted for flotation.

(&) The Feeding of Seivage.

The presence of Oscillatoria may be taken as an index of the
amount of sewage in the food. It also shows a well-defined
gradation in the several feeding stations: (i) no Oscillatoria
was recorded from the experimental animals above the sewer
outlet; (2) very little appeared in the checks kept in the un-
screened stream at this point; (3) somewhat more Oscillatoria,
consisting of very incomplete filaments, was seen in the experi-
mental animals below the sewer ; and (4) the checks below the
sewer contained numerous large fragments of Oscillatoria, and
many relatively large bits of debris never met with elsewhere.

Mussels which had been kept in the mouth of the creek for
several weeeks prior to these experiments were opened at the
same time. The styles were well formed and without color. On
various occasions the styles of the experimental animals reacted
differently. At one time all of the styles were whitish when par-
tially renewed, having a distinct white, spiral core. In those most
perfectly formed the white color was disappearing, and the entire
mass was becoming more solid and more hyaline. On other occa-
sions the newly formed styles were of the typical amber hue
which suggested diatomin. Subsequently mussels opened here
contained sometimes whitish, sometimes colorless styles. At
times freshly formed styles were found which were green in the

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 221

stomach and becoming white in the style sac, indicating a change
from one color to the other.

Only one explanation of the above differences offers itself—
namely the effect of bacteria. These, through a mass effect or
through the breaking down of the style substance itself, are prob-
ably responsible for the white color. Rice (I.e.) has shown that
the abundance of bacteria and of nitrates is here subject to pro-
found variation, on account of the periodic discharge from the
septic tank mentioned above. Whatever the cause of the white
color it might have been expected to be mingled with green from
the normal food brought down from above. That green is actu-
ally present in concealment is shown by the boiling of such styles
(p. 226). The white style pertains mostly to the creek. In the
lake a white style is observed only during low water in mussels
which have been feeding near the creek outlet. During freshets
and great dilution of the bacteria it does not occur even in the
creek. These observations may be taken as a further indication
of the direct dependence of the crystalline style upon the char-
acter of the food.

(c) The Feeding of Infusions.

Previously I had observed the renewal of the style and the
accumulation of material in the alimentary tract of animals fed
with hay infusions rich in ciliates (Allen, I.e.). The unmistak-
able finding of protozoan fragments in the stomach showed that
some such material is ingested, and I was satisfied that the style
renewal was due to their presence. The above nannoplankton
studies suggested that there might also be food value in the bac-
teria and flagellates which are present in such concentration in
infusions. The former experiments upon the feeding of infu-
sions were repeated with the same results. Again, white crystal-
line styles appeared (p. 220). However, in order to determine if
the bacteria and flagellates present may be responsible for the
style renewal an attempt was made to separate them from the
large ciliates. The above method for the separation of plankton
was applied here. It was possible, at any rate, to dispose of most
of the bacteria by washing. The process resulted in the death
of many ciliates due to crushing and to the change into fresher

222 WILLIAM RAY ALLEN.

water. However, it was possible to accumulate a considerable
mass of living and fragmented infusoria. Starved mussels fed
upon this concentrate did not renew the style. This negative
result must not be trusted too implicitly, considering that, as re-
ported above, ciliate material is sometimes found in the stomach.
Yet the present experiment points toward a greater utilization of
the smaller organisms of infusions than of relatively large ciliates.
Aside from the ciliated protozoa the infusorian population con-
sisted mostly of extremely minute forms. The maceration ex-
periments described elsewhere (p. 228) show conclusively that the
Unionid gill is capable of intercepting very small material in-
deed— even the pyrenoids of algae.

It may be questioned whether the formation of a crystalline
style is a reliable index of the feeding activity. In my opinion it
is as reliable as a direct examination of the alimentary tract. On
some occasions food may be found in the intestine before regen-
eration of the style is perceptible. But, on the contrary, there
are as many occasions when a starved mussel recently fed is seen
to have the beginnings of a style before anything is readily recov-
erable from the intestine. In far the greater number of cases
examined the synchronism between style renewal and the pres-
ence of food is exact. The bearings of this upon the significance
of the style are discussed below (p. 229).

Resume. — The more minute plankton organisms are of as great
nutrient value as the more conspicuous, often undigested matter
commonly listed from rectal or fecal examinations. However
the net plankton is shown to have a food value as well. Experi-
ments which more or less perfectly separated the net- from nan-
noplankton show that both are capable of re-forming the crystal-
line styles of starved mussels. The minute flagellates sometimes
exceed the volume of net plankton in lakes many fold. Since it
is certain that they can be entrapped by the gills of the mussel
and can be ingested, it is likely that their rarity in the rectum is
due to the fact that most of them have been digested. Infusions
which have nothing of food value except ciliates and minute or-
ganisms renew the style.

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 223

4. THE PHYSIOLOGY OF THE CRYSTALLINE STYLE.

Everyone, myself included, who has dealt with the crystalline
style during the past decade, has made apology for adding to the
bulky list of the things not certainly known concerning that
organ.

(a) Recent Studies.

The most thoroughgoing and satisfactory account of the crys-
talline style is that of Nelson ('18). He has assembled and or-
ganized the literature on the subject to the minutest detail, has
very effectualy eliminated most of the groundless speculations,
and has sifted out the truth contained in the rest. Nelson's work
on the morphology is of a sort which virtually closes that sub-
ject. All future studies of the crystalline style may well make
Nelson's work the point of departure.2 There is no occasion to
review the literature here or to describe either the style itself or
the associated portions of the alimentary tract. I shall be content
to record the physiological data which have accumulated during
the intermittent observations made over a period of some six
years.

Most of the writers, with the exception of Mitra ('01), and
Nelson (l.c.~), have taken a viewpoint which has been fatally
erroneous, it seems to me. Despite all the divergent speculations
which observers have permitted themselves to make (a point well
reviewed by Nelson), they have really been looking for one
explanation — the most plausible function that this organ might
be supposed to perform. Few have granted the probability that
two or more uses of the style might exist concurrently.

- The preceding sentence, when written, was prophetic. For just as this
paper is about ready for the press, Edmondson's timely account of the crystal-
line style in Mya arenaria has appeared ('20). Like Nelson, he has devoted
considerable study to the morphology, but has centered his attention upon the
renewal of the style. It is gratifying to find others interested in the physio-
logical study of the style, for, aside from its chemistry, most work has been
done from the viewpoint of structure.

This author's account of Mya arenaria shows the style to have diverged
very far, indeed, from its homologue in the Unionidze. It lies in a distinct
caecum. Operative methods instead of starvation were necessary to remove
it. It is a very solid structure, nearly insoluble, and nearly devoid of albu-
minoids. Its regeneration in Mya precedes rather than follows resumption of
feeding. Seventy-four days were required for its reformation in Mya, while
one day more or less suffices in the Unionidx.

224 WILLIAM RAY ALLEN!

Mitra (I.e.) was the first to recognize clearly that the crystal-
line style may meet several needs. The fact that it is dissolved
when food is wanting, and that in solution it may be taken up by
the blood, leads to the conclusion that it is (so far as it goes) a
reserve of nutriment. Furthermore, the work of Mitra and
others has shown that it bears enzymes capable of furthering
starch digestion. Dr. Scott Edwards has kindly checked over
this matter for me, with confirmatory results. Insofar as suitable
food is brought into contact with it, it is a means of supplying
digestive ferments.

Nelson (I.e.) has shown very well that the rotation of the
crystalline style against the " gastric shield "in the stomach shreds
the dissolving end so as to form a brush. The rotation of this
brush sweeps the food into the proper ciliated channels, aids in
dissolving the food out of the mucous masses in which it reaches
the stomach, and acts as a substitute for peristalsis in mingling
food with the digestive fluids from the liver, etc. It might have
been pointed out that these movements afford a ready means of
bringing the contained enzymes of the style into thorough contact
with the food.

Since the style actually accomplishes all these things, we can
not choose any one of them as the function for which it was
designed. We cannot assert that the style is exactly adapted to
perform any one of them, or that it is the function which it has
always performed in ancestral forms. If such were the case one
might expect to find somewhere in the more primitive existing
species a style little changed from the ancestral condition. But
the Najades have taken to fresh water, and as a result have be-
come profoundly modified in life history, ontogeny, and structure.
While the crystalline style has shown as little structural change as
any organ, it is not improbable that its relation to the organism
as a whole, to metabolism, has suffered changes of which we know
nothing.

Nelson suggests a further function of the style sac, that of
returning undigested material from the intestine to the stomach,
to prevent the loss of food. In Modiohts, which undergoes a
periodic cessation of feeding and dissolution of the style, the

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 225

cilia of the sac periodically carry a stream of food from the
intestine through it into the stomach. In common with all
ciliated epithelia this organ raises the very interesting and equally
difficult problem of accounting for the present direction of the
beat of the cilia. In the simpler ancestral forms the beat most
probably was in the opposite direction.

The style sac of Modiolus contains more food in winter than in
summer. It is difficult to see why the structure in Modiolus
should be (as Nelson reports) more effective in winter, when the
metabolism is low and the food requirement slight, than in
summer, when the demand for food is greater. In winter the
secretion of the style substance is slowed down by the tempera-
ture to such an extent that the organ is not promptly re-formed
with each feeding period (p. 229). The style sac therefore con-
tains no style and may be utilized to reconvey food from the
intestine to the stomach. In the Unionidse it is certain that the
return of particles through the style sac is a phenomenon which
takes place normally only after starvation. Of the many hun-
dreds of specimens examined when taken from the water, not half
a dozen were ever found in which the style was lacking.

Usually the newly formed style has an abnormally large core
of plankton. This indicates that the first undigested or partially
digested material which streams into the intestine is diverted at
the posterior end of the style sac and carried forward again into
the stomach. In the meantime the glands of the typhlosole
(Nelson, l.c.~) begin secreting and wrapping the spiral of the
style substance about this core. The streaming in of materials
from the intestine is limited more and more as the style more and
more completely fills the lumen of the sac. After the style moves
forward and has been dissolved its entire length, the newer por-
tion, with a diminished core, has entirely replaced the original
portion with its loose structure and large core of food. This
takes place twice daily in Modiolus or Ostrcca, and the newly re-
generated style is thus oftener encountered. In Lampsilis or
Anodonta under normal circumstances, it is a very infrequent
occurrence. My observation has been to a very great extent upon
lake forms, which are not subject to many vicissitudes. It is

226 WILLIAM RAY ALLEN.

presumable that river forms, if taken at the right times, demon-
strate a more periodic activity in response to the rise and fall of
the current, the degree of turbidity, etc.

Nelson is probably in accord with these reservations concerning
the Unionidse, for he agrees (I.e., p. 100) that the formation of
the style is directly dependent upon the food supply. In the
marine forms the return of food through the style sac to the
stomach is considerable. In the Unionids the return of food
through the style sac has become reduced because the sac is
usually occupied by the style.

The spiral character of the style, caused, as demonstrated by
Nelson, by its axial rotation, has been brought out nicely in three
ways in my observation :

fa) A regenerated style in one starved Anodonta was found
to have a great deal of green matter throughout, and but a very
slight amount of style substance. It had grown to at least normal
diameter. When kept for a time the secreted portion dissolved
out, leaving the green cohering portion wrapped about the core
like the threads of a screw, or a lathe shaving.

(&) A few whitish styles were boiled for a short time in strong
sodium hydroxide. They were much reduced, and the residue
was in the form of a close rope-like spiral.3 The white appear-
ance had given place to the amber-green color characteristic of
most newly regenerated styles.

(c ) The actual rotation of the style has often been observed.

Examinations were frequently made of the food substances
which could be recovered from the core of more or less com-
pletely regenerated styles. The food never showed a perceptibly
greater degree of digestion at the stomach end than posteriorly.
It is not likely, therefore, that the digestive processes are much
furthered during progress through the style or style sac toward
the stomach.

(b) The Feeding of Specific Substances.

A series of experiments in feeding certain substances, and in
forced feeding by injection into the stomach, were undertaken

3 Edmondson's ('20) figures of partially formed styles very well represent
these.

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 22 7

to throw light upon the actual stimulus which initiates the renewal
of the style. The stimuli may be either mechanical or chemical.

Carborundum, carmine, starch, etc., of varying fineness, were
introduced into the incurrent siphon with the streaming water.
Such organic or inorganic material, however neutral, of whatso-
ever dilution, or however administered through the respiratory
water, were never found subsequently in the alimentary canal.
Nor did the molluscs ever display any indication of style renewal
in response to these things. The experiment has a further
significance to be discussed on page 229.

A culture of Glccocapsa was looked over carefully and found
to have very few organisms of the size of Glccocapsa, but much
coarse debris. This material was washed into a jar with active,
starved mussels, and agitated from time to time to prevent its
settling. After eighteen hours the mussels were examined. The
crystalline styles were partially renewed in all cases. In others
kept in jars of Glccocapsa not stirred frequently the styles failed
to show any indications of re-forming, even after three or four
days. Evidently not enough food to stimulate style formation
had been taken into the siphons. In all, small masses of Glccocapsa
were encountered in the rectum, in the stomach, and in clots of
mucus upon the gills and palps. The clots were almost pure
Glccocapsa. The stomach and intestine contained minute frag-
ments of green, partially digested individuals, and sometimes
Glccocapsa cells without the capsule. There is thus no doubt but
that a pretty rigid selection of the alga from the coarser matter
with it was taking place, and that the alga was being digested.

The frequent occurrence of the Glceocapsa in the rectal con-
tents, still wrapped in mucus, shows the effect of the want of
the style. Had that organ been present the mucus masses would
probably have been torn up, the alga freed in the stomach and
exposed to digestive action. It appears that the process of di-
gestion does not function perfectly, even prior to the formation
of a style, and not even a hungry mussel exposes all particles
equally well to the digestive fluids.4

4 Edmondson finds the alimentary tract of Mya arenarla empty of food
until the style is partially replaced.

228 WILLIAM RAY ALLEN.

A quantity of Spirogyra and other filamentous algae was cut
as finely as possible with scissors, then macerated with a pestle.
Examined microscopically there were found fragments of cell-
wall varying in size, fragmented chloroplasts, and pyrenoids.
Without screening, this material was fed, in considerable quantity
to starved mussels. The water was agitated occasionally so as
to keep some of the macerated material in suspension. One mass
of alga had been taken from a dish in which decomposition had
gone far. Although the mussels held the siphons nearly closed
in the decomposing culture, they nevertheless ingested sufficient
alga to answer the purpose of the experiment. Feeding the de-
composed alga is comparable to the feeding of infusions, and the
animals reacted similarly. In all cases of feeding infusion, de-
caying alga, and the feeding of mussels below the sewer outlet in
Pocahontas creek, the regenerated crystalline styles had the same
milk-white appearance. Thus there is no doubt that in all cases
the color was due to bacteria. Mussels fed upon macerated
fresh Zygnenia had clear, and colorless or green, styles.

In one mussel fed upon a maceration of Spirogyra the partially
regenerated style had a large core, and only two or three thin
layers of style substance. This gave it the proportions and ap-
pearance of a rubber tube. When stretched out in a watch crystal
and placed under the weight of a cover'glass, the contents slowly
oozed out at a broken point. The core was then seen to consist
almost entirely of pyrenoids (or like bodies) closely packed, and
in very great quantity. Their mass had a gray-yellow color.
Only here and there was there a minute spot of green. Not a
trace of cell walls or of a spiral fragment of chloroplast was
found here or in the stomach. We have then another case of the
rigid selection of food particles, and a little suggestion of the
character of the materials which are capable of inciting the secre-
tion of a new style. Again we have evidence that the gill is
capable of taking very small particles from the water.

In a decaying macerated Spirogyra culture several starved
mussels kept the siphons almost entirely closed. When a few c.c.
of alcohol were added they shortly began and continued siphoning
vigorously. Yet the food sorting mechanism functioned normal-

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 22Q

ly, for no stomachs nor intestines were found to contain larger
fragments than usually occur there.

(c) Forced Feeding.

From the above experiments, and on grounds discussed else-
where (p. 227), it is seen that the ingesting apparatus exercises
considerable choice, and that (at least under experimental condi-
tions) only certain sorts of material are admitted to the stomach.

An attempt was made to introduce distasteful matter into a
starved mussel with the food. When fed alone, carmine had
never been ingested. It was therefore administered with Glcco-
capsa and macerated Spirogyra. In no case was it found in the
alimentary canal. Very little of the food entered, for thai matter.
The presence of the carmine caused a rejection of most of the
food as well.

In order to ascertain if substances rejected by the mouth might
yet have the power to stimulate the secretion of the crystalline
style, these were introduced little by little through a fine pipette
directly into the stomach. Fine carborundum (120-180 gauge)
carmine, and starch were tried. In none of these cases was any
trace of a style to be found later. It should be explained that the
shock of operation was not alone responsible for this failure, for
when Glococapsa was fed to the animals in the same way, it was
capable of renewing the style to a slight degree. There is suffi-
cient ground for the conclusion that mere mechanical stimulation
of the intestine or style sac on the part of fine particles is not
sufficient to initiate the formation of the style. There must be a
stimulus of a chemical nature as well. A reaction to the feeding-
activity might have incited secretion through reflexes from the
palps. Yet this is inadequate to account for the renewal of the
style when Glccocapsa was administered through the stomach wall.

(d) The Effect of Temperature on Style Renewal.

With the approach of winter it becomes more and more diffi-
cult to secure a prompt renewal of the crystalline style on the
resumption of feeding. Where experiments are made in water
of quite low temperature the same behavior is observed. Riddle,

230 WILLIAM RAY ALLEN.

('09), Krogh ('14) and others have shown that the rate of meta-
bolic processes is related to body temperature. In this case the
decrease in temperature probably directly affects the rate of secre-
tion of the typhlosole glands. It is not improbable that a seasonal
metabolic rhythm exists.5

In order to test the effect of temperature alone in this matter
and to eliminate the other possible elements, the following experi-
ment was carried out :

Checks were kept at room temperature. In each repetition of
the experiment one jar was placed in the cold water of the outlet
of an artesian well. When the temperature was sufficiently re-
duced the experimental starved animals were introduced into
their respective jars. Meantime concentrations of lake plankton
were made by filtering through fine bolting-silk, and the plankton
remaining in the bucket washed out into the check and experi-
mental jars. Each mussel had a jar to itself. The water in all
had a decided green tint when agitated, for the concentration was
in all cases from 100 volumes to i. The jars were well shaded
to eliminate the possible action of sunlight in orienting the plank-
tonts, and thus keeping the food equally available to all. The
amount of water was equalized, and the mussels so placed that the
exhalent stream played obliquely upon the sides of the jar and
maintained an eddy to keep the planktonts in circulation.

The material was collected at about 10:00 A.M.. and about
three hours were allowed for the temperature adjustment. The
experiments began at about i :oo P.M., and continued from three
to four and one-half hours, usually four. The average tempera-
ture of the experimental jars at the beginning of the experiments
was 13.4° C, varying between 12.6° C. and 14.0° C. The aver-
age of the same jars at the end of the experiments was 13.1° C.,
varying between 12.0° C. and 14.4° C. There was an average
loss in temperature of 0.3° in these jars. The checks gave an
average at the beginning of 26.0° C., varying between 23.6° C.
and 30.0° ; at the end an average of 27.2° C., and a variation
between 24.0° and 29.6° C. The average rise in temperature of

5 The crystalline style of Mya arenaria is reformed more rapidly during
summer than winter. Edmondson (I. c.) ascribes this to the increased metab-
olism of the approaching breeding season.

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 23!

the checks was 1.2° C. At the beginning the checks averaged
12.6° higher than the others, and at the end averaged 14.1° higher.
Nearly ten degrees (9.6) separated the lowest check from the
highest temperature in an experimental jar. As was to be ex-
pected the atmospheric temperatures created considerable varia-
tion in the checks (6.4°) and much less in the cooled jars
(2.4° C.)

Of the twenty animals used only ten showed partially renewed
styles. Of these ten only two occurred at the reduced tempera-
ture, and eight in the checks. The two which appeared in the
low temperature were smaller than any of the eight formed in the
checks. Moreover, the two were both in Anodonta, and Ano-
donta has shown a greater response always than Lain[>silis, never
failing to show some renewal. Lampsilis more slowly loses and
more slowly regains its style than Anodonta.

While the number used is small, exact quantitative results are
here unnecessary, and there is sufficient demonstration of the
qualitative effect of temperature upon style formation. This
effect may be partially explainable through the effect of tempera-
ture upon the cilia and the rate of ingestion. But the reason men-
tioned above is probably more pertinent, for the intestine was
usually found to contain food in the experimental animals.

So long as the quality of the ingested material is right, the
quantity required to initiate the formation of the style is very
small. At times a single battery jar of water dipped at random
from the surface of the littoral has contained sufficient food to
restore it, in part. Held to the light the water had given no hint
of green. But after it had been siphoned from one to two hours,
the resulting thread-like crystalline style contained a conspicuous
core of green.

Where food is abundant the length of time needed to renew
secretion after the beginning of feeding is very short. A fair
beginning may sometimes be observed within fifteen to thirty
minutes. Large well-formed styles are sometimes secreted in
four hours or less. The time depends largely upon the degree of
.starvation. More often twenty-four hours, at least, are required.
On the whole it is a much more deliberate process than in some

232 WILLIAM RAY ALLEN.

tidal forms, where the breaking down and renewal of the style
occur rhythmically.

Of passing interest is the observation that small, newly formed
styles sometimes may be seen coiled up in the stomach, where they
have pushed forth more rapidly than they could be broken up
and dissolved against the gastric shield.

Resume. — It is here contended that the crystalline style accom-
plishes a number of purposes, for none of which it is entirely
indispensable, nor entirely a perfect adaptation ; that it is no longer
performing an identical, single, primitive function traceable to a
primitive Lamellibranch ancestor. The response of this organ
to similar conditions is much the same in various bivalves ; but
the tranquil life of the lake has stabilized the feeding activity and
the style formation in the Unionids, while the styles of some
species inhabiting the marine littoral are profoundly affected by
the tidal phenomena.

The formation and dissolution of the crystalline style goes
on in the same way that a paper, candle-lighter might, if extended
to an indefinite length by rolling up a sheet of paper of indefinite
length, and burning off the free end as rapidly as new paper is
added at the other.

There is no evidence that digestion is furthered during the
passage of food through the style or style sac.

The feeding of inert substances, both normally and through
the stomach wall, indicates that the mechanical stimulation of the
wall of the enteron is not alone the cause of the secretion of a
new style. The rate of formation of the style is shown to depend
in part upon temperature. Little food and little time are required
to set the process going.

5. THE MECHANISM OF INGESTION.

It has long been known that the gills with their great multiplica-
tion of surface are responsible for the movement of respiratory
water, and for the concentration of food material from the

»

water. It is surprising to encounter in the work of so eminent a
student as Simpson, written only a score of years ago ('99), a
statement that the siphoning is due to the waving of the palps.

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 233

It was shown by Posner ('75), Wallengren ('05), and other?,
against the contention of M'Alpine ('88), that the collections of
food are transmitted to the labial palps, and by their cilia to the
mouth. The writer shows (I.e.] that the ciliary streams of the
upper portion of the mantle chamber all tend toward the mouth;
while those of the lower portion of the visceral mass and mantle
lead away from the mouth. The latter accomplish the duty
ascribed by M'Alpine to all the cilia, that of freeing the mantle
chamber of heavier materials and rejected food clots.

It was stated by the writer that the food material is subject to
rejection at four points: (i) the siphons, (2) the point on gills
and mantle where the food stream passes to the palps, (3) the
furrowed surfaces of the palps, and (4) the lips, at the mouth.
More recent observations have all corroborated this. Perhaps
the fact has not been sufficiently stressed that only an unusual
chemical or tactual stimulus results in the closure of the siphons
or lips. The palps somewhat oftener refuse masses from the
gills and mantle by turning aside. The greater number of reac-
tions occur as the food stream passes between the contiguous
palp surfaces.

The work of Wallengren ('05), and others, has demonstrated
the action of the labial palps, the ridges of which are capable of
reversing the food stream. Near the distal margin of each ridge
of the palp surface the beat of the cilia is toward the apex, both
in front and behind. When the ridges are inclined forward, the
effective beat of the cilia is forward ; when the ridges alter their
axis, the backward-beating cilia are brought into play and the
others turned under.

The course of the ciliary streams at the bottoms of the furrows
between the ridges is much more difficult to observe. Wallengren
believes that the cilia strike downwards to the edge of the palps,
and that the resulting streams belong to the excurrent system.
Siebert ('13), working on Anodonta ccllensis, says they strike
in the opposite direction — upward to the apex of the inverted V
formed by the two palps, thence forward to the lips and mouth.

234 WILLIAM RAY ALLEN.

(a) The Role of the Labial Palps.

The writer has checked the matter as carefully as possible, and
believes that there is ground for the views of both Wallengren
and Siebert. The details of arrangement may not correspond
exactly in the several species. On the lower half of the palps the
cilia under consideration usually strike downward, and those of
the tipper half strike upward. Thus the lighter and finer particles
tend to be drawn upward and forward as food, while the heavier,
coarser, materials are more likely to be carried downward. The
differentiation of the mechanism corresponds pretty well to that
of the upper and lower portions of the ciliated surfaces of the
mantle chamber as a whole. It is impossible to make direct
observation of the streaming on any one ridge of the palps. But
where substances of varying fineness are placed together on the
palps, such as carborundum dust and carmine, there is a tendency
to assort them. The carborundum particles move along the
apices of the ridges and are carried nearly lengthwise of the palps.
The carmine gravitates farther into the furrows between the
ridges. Near the lower margin of the palps carmine is carried
obliquely downward and forward, and on the average reaches the
lower edge of the palps before the carborundum. When placed
on the upper portion of the contiguous palp surface, carmine is
drawn upward and forward to a greater extent than the car-
borundum, then forward toward the mouth.

The respective upward or downward pull upon the carmine may
be accentuated by stretching the palps lengthwise, thus drawing
the ridges farther apart and exposing the cilia of the furrows to
a greater extent.

Attempts were made to effect a reversal of the ciliary currents
of the furrowed surface of the palps by injections of curari,
strychnin, atropin, pilocarpine, and by electric stimulation. The
injections were made through the body wall into the sinuses near
the base of the palps. Observations were made at various in-
tervals from ten minutes up to several hours after injection. It
was never found possible to control the reaction. There was a
perceptible response to none of the several drugs except strychnin.
This sometimes caused a contraction, at other times a relaxation

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 235

of the palps. When the palps were in a state of contraction, most
of the streaming carmine and carborundum were drawn to the
edge and into the mantle chamber. These mieager results tend to
corroborate previous observations on the reaction of the palps-
that food may be rejected through a greater or less erection of the
ridges. Negative results prove nothing, while ever so slight posi-
tive evidence may be taken as an indication that a reversal of the
ciliary streams can, and actually does, take place, through the
bringing of another set of cilia into play.

The above shows clearly that the palps bear two sets of cilia
working at right angles to each other. Due to their interaction
materials often travel obliquely downward or obliquely upward.
The former materials are eliminated at the lower edge of the
palps, the latter reach the mouth.

The effectiveness of the assorting mechanism is well brought
out by the measurements of the ingested particles. The largest
fragment I have ever secured from the enteron was found in the
intestine — a pinnately branched alga 3.3 mm. in length, probably
Mv.roncina. The second largest particle was a bit of Oscillatoria,
1.5 mm. in length. It is unusual in Winona Lake mussels to en-
counter fragments of greater length than 500 microns. Starved
individuals, probably experiencing a sensation akin to hunger, are
observed to ingest freely much more large material than under
normal circumstances.

Nelson ('18) has described the action of the food sorting
caecum of Modiolns, a diverticuluin of the stomach. Since
Modiolus ingests large quantities of sand in its periodic feeding,
Nelson's ccccitin affords a means of separating food from sand.
It has been shown elsewhere in this paper that feeding in the
Unionidse is a more constant function, and that little sand and
mud are taken into the stomach. The gills and labial palps are an
entirely sufficient assorting mechanism.

(b) The Gills as an Assorting Mechanism.

Little attention has been given to the gills as having a possible
food-sorting function. I find that they play no small part. Clots
of mucus taken from various parts of the gills and palps have

236 WILLIAM RAY ALLEN.

been examined, and often contain little but the finest ingestible
material. This was well shown in the case of a Glccocapsa cul-
ture fed to a starved mussel. The culture was very pure except
for numerous fragments much coarser than the Glccocapsa itself.
There was an almost complete separation of Glccocapsa from the
other material by the gills themselves. Almost all the larger
fragments had been separated out by the gills themselves before
the finer had been agglutinated in mucus. Little of the former
was found in the masses present in the alimentary tract.

The marsupial function of the gravid gill of the female inter-
feres somewhat with its respiratory and food collecting functions.
Ortmann ('12) has shown that secondary water tubes appear, in
which water circulates about the egg masses, and accomplishes
the aeration of the eggs and glochidia. Yet the volume of water
siphoned is much less than in the case of the non-marsupial gills.
This is well brought out by the fact that the gravid females almost
invariably regenerate the crystalline style much more slowly than
others. When kept under artificial conditions for some time the
gill-masses are usually aborted, another indication that the gravid
gills are unable to meet all the demand upon them. The greater
remoteness of the marsupial gill has suggested that it has become
differentiated for the storage of the eggs and has lost its food
collecting function. This notion is pretty well refuted by the
facts mentioned above concerning the regeneration of the style in
gravid females as compared with non-gravid females. It has
also been suggested that the mantle has taken over much of the
respiratory duty of the gills. If this were true the gravid female
should be under no special respiratory difficulty. When first
brought into captivity these females die at a much greater rate
tlfan others. The accessory water-tubes seem to be a very im-
perfect makeshift, sufficient perhaps for the glochidia, but afford-
ing the mother little aid.

Since in the gills and palps there exists a mechanism well
adapted for the sorting of food ; since both observationally and
experimentally this mechanism is shown to accomplish a con-
centration of food ; and since the contents of the alimentary
canal have a decided green or brown color due to such concentra-

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 237

tion, we may feel safe in the reiterated conclusion that the
Unionids exercise choice in the ingestion of materials. As stated
by Zacharias ('07), Petersen ('n), the writer ('14), Baker
('16), and others, considerable quantities of inorganic and organic
debris are carried into the stomach with food. Probably much
of the stuff which Evermann and Clark ('17) call "mud" is
organic. The fact that neither they nor other writers list sand in
the stomach contents is further evidence of a selection of food
material, and that river species are not an exception. Starved
mussels were placed in the lake in two localities — (i) an open
leeward shore in clear water; (2) near the outlet of Pocahontas
creek, in muddy water following a rainstorm. The mussels of the
first situation reformed the crystalline styles within a few hours.
The others contained great quantities of muddy mucus, and did
not have well renewed styles until the following day. The slow
renewal of the style may be accounted for in part by the dilution
of the food. But the presence of mud must be held partially ac-
countable, for immediately after the subsidence and clearing of
the water it was always found to contain ample food material to
renew the style promptly.

In most species the position of the siphons at some distance
above the substratum tends to keep out most of the grosser
particles, admitting little but plankton and other materials in
suspension.

In the discussion of the crystalline style (p. 227) the feeding of
specific inert substances were recounted. When such materials
which were readily identifiable were admitted with the incurrent
water they were in no case found in the alimentary tract. As
far as size is concerned these particles could very readily have
entered the mouth. Since all were rigidly excluded we cannot
doubt that sense organs exist for their detection, and that the
assorting mechanism is a fairly effective one. Of the materials
mentioned only starch might be expected to have a food value,
though we cannot assume that it is in acceptable form. As a
matter of fact the rejecting reactions were more vigorous in re-
sponse to starch than to the other substances.

In the above experiments on the crystalline style neutral sub-

238 WILLIAM RAY ALLEN.

stances were introduced through the body wall into the stomach.
At later periods the intestine and rectum were opened. Carmine
and starch grains were recognized throughout the length of the
alimentary canal. Only a few carborundum flakes were found
in the intestine, and the rest were not carried out of the stomach.
It is thus shown that the ciliary streams of the intestine are
capable of manipulating only minute particles. The cilia are too
small or too sparse to take care of the I2o-gauge carborundum,
even in suspension in the liquid of the gut. Thus they must be
altogether inadequate to keep a stream of sand in motion, if sand
were ingested, unless it were of extremely fine grade.

(c) The Marly Incrustation of the Shell.

Since the dense, limy incrustation deposited on the exposed por-
tions of the shells of lake mussels is the site of the active prolifera-
tion of diatoms, I suggested (I.e.') that this might be a source of
food. In order to test its food value the following simple experi-
ment was made :

A number of freshly collected mussels were placed in an aquar-
ium ; an equal number, having the incrustation scraped and
washed off, acted as controls. During intervals, covering several
days the animals were opened and the condition of the crystalline
styles noted. The experimental animals were found to contain a
trace of the style up to the fifth day, while the checks had virtually
lost it by the end of the second day and entirely lost it on the
third. Of course the crowding created rather special conditions,
unlike those of the lake. The conclusion is that the incrustation
contains considerable food.

6. SUMMARY AND CONCLUSIONS.

1. Feeding in the Najades is a nearly constant function under
normal conditions. The presence of much undigested and some-
times living matter in the rectum and feces shows that there is a
greater fluctuation in the degree of digestion than in the rate of
ingestion.

2. The posture of a mussel has no effect upon the continuity
of the feeding process, a further indication that under normal

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 239

circumstances ingestion may go on with less effort than an inter-
ruption of feeding.

3. The return of undigested material from the intestine through
the style sac to the stomach is an unusual occurrence in the
Najades, which takes place only after periods of starvation, and
which is interrupted With the reformation of the style. It is a
function much less significant in the Najades than in the tidal
forms.

4. Experiments in feeding relatively finer and coarser plankton
show that both are capable at least of stimulating the renewal of
the crystalline style. Both have food value. It is probable that
the nannoplankton furnishes a much greater part of a mussel's
food than has been suspected. The studies of intestinal contents
of the Unionids have not demonstrated what the actual food is,
but rather the undigested residue. Experiment here has shown,
however, that the organisms undigested in the feces are some-
times digested, under another set of conditions.

5. Starved animals fed in Pocahontas creek below the outlet of
a sewer showed the following peculiarities in intestinal contents :
( i ) the occurrence of many relatively large organic fragments ;
(2) abundance of minute flagellates; and (3) great quantities of
Oscillatoria filament.

6. The regeneration of the crystalline style is in response to the
ingestion of food, and not due to the physico-chemical character
of the water.

7. Creek-fed mussels show a variation in the color of the style
through hyaline, amber, and milky. The apparently rhythmic
character of this variation corresponds roughly to the variation of
sewage discharged from a septic tank. The milky color is ac-
counted for by the presence of bacteria.

8. A repetition of experiments in feeding infusions indicates
that flagellates (and bacteria) present are responsible to a greater
extent for the renewal of the style than are the bulkier ciliates.

9. The reformation of the crystalline style is a satisfactory
index of the renewal of the feeding activity.

10. The function of the crystalline style varies. It may more
or less imperfectly perform several functions at once. The

240 WILLIAM RAY ALLEN.

rhythmic loss and renewal of the style in tidal forms has no
parallel in freshwater species.

11. The feeding of specific substances in high concentration
never produces a renewal of the crystalline style unless such sub-
tances have a food value. No indication was observed that the
stimulus for its secretion is a mechanical one.

12. The style is much less readily formed in autumn or winter.
That temperature is responsible is shown by experiments in which
starved mussels were fed a concentration of plankton in water of
high and low temperature, respectively.

13. A very small amount of plankton is sufficient to stimulate
style formation. Also only a short time is required.

14. The labial palps are the primary assorting mechanism.
The gills are of considerable importance in this matter also.

15. Sand is never ingested, at least by lake forms, and mud but
slightly. Much less inorganic debris finds its way into the
stomach than would be the case if selection were not exercised bv
the gills and palps.

1 6. The cilia of the alimentary canal are unable to move coarse
materials, or to maintain a stream of sand or heavy mud.

17. Gravidity of the gills is a serious hindrance to the respira-
tory and alimentary activities.

LITERATURE CITED.
Allen, W. R.

'14 The Food and Feeding Habits of Freshwater Mussels. BIOL. BULL.,

Vol. 27, pp. 127-139.
Baker, F. C.

'16 The Relation of Mollusks to Fish in Oneida Lake. Tech. Pub. no. 4,

N. Y. State Coll. Forestry, pp. 1-366.
Baker, F. C.

'18 The Productivity of Invertebrate Fish Food on the Bottom of Oneida
Lake, with Special Reference to Mollusks. Tech. Pub. no. 9, N. Y.
State Coll. Forestry, pp. 1-264.
Edmondson, C. H.

'20 The Reformation of the Crystalline Style in Mya arenaria after Ex-
traction. Jour. Exp. Zool., Vol. 30, pp. 259-291.
Ege, R., and Krogh, A.

'14 On the Relation between the Temperature and the Respiratory Ex-
change in Fishes. Int. Rev. ges. Hydrobiol., Bd. 7, pp. 48-55.
Evermann, B. W., and Clark, H. W.

'17 The Unionidae of Lake Maxinkuckee. Proc. Ind. Acad. Sci., pp. 251-
285.

STUDIES ON BIOLOGY OF FRESHWATER MUSSELS. 24!

Headlee, T. J., and Simonton, J.

'03 Ecological Notes on the Mussels of Winona Lake. Proc. Ind. Acad.

Sci., pp. I73-I79-
Lohmann, H.

'n Uber das Nannoplankton und die Zentrifugierung kleinster Wasser-
proben zur Gewinnung desselben in Lebenden Zustande. Int. Revue
ges. Hydrobiol., Bd. 4, pp. 1—38.
Nelson, T. C.

'18 On the Origin, Nature, and Function of the Crystalline Style of Lamel-
libranchs. Jour. Morph., Vol. 31, pp. 53-111. Has an exhaustive
bibliography on the crystalline style.
Ortmann, A. E.

'12 Notes upon the Families and Genera of the Najades. An. Carnegie

Mus., Vol. 8, pp. 221—365.
Petersen, C. G. J., and Boysen-Jensen, P.

'n The Valuation of the Sea. Kept. Dan. Biol. Sta., Vol. 20.
Rice, T. B.

'16 A Study of the Relations between Plant Growth and Combined Nitro-
gen in Winona Lake. Proc. Ind. Acad. Sci., pp. 333-362.
Riddle, 0.

'09 The Rate of Digestion in Cold-blooded Vertebrates — the Influence of

Season and Temperature. Am. Jour. Physiol., Vol. 24, No. 5.
Simpson, C. T.

'oo Synopsis of the Naiades, or Pearly Fresh-water Mussels. Proc. U. S.

Nat. Mus., Vol. 22, pp. 501-1044.
Ward, H. B., and Whipple, G. C.

'18 Freshwater Biology. New York, 1111 pp., and 1547 figs.
Zacharias, 0.

'07 Planktonalgen als Molluskennahrung. Archiv f. Hydrobiologie u.
Planktonkunde, Bd. 2, pp. 358-361.

Vol. XL. May, 1921. No. 5

BIOLOGICAL BULLETIN

INTER-PERIODIC CORRELATION IN THE ANALYSIS

OF GROWTH.

J. ARTHUR HARRIS AND H. S. REED,
STATION FOR EXPERIMENTAL EVOLUTION, COLD SPRING HARBOR, LONG ISLAND.

I. INTRODUCTORY.

In the literature of growth, mathematical equations to describe
changes in the actual size of the organism, or changes in the
growth rate, are finding continuously widening applications. One
has merely to refer to the papers by Robertson, Miyake, Moeser,
Ostwald, Reed and Holland (1919), and Reed (I92O)1 for illus-
trations.

The criticism usually directed against such work is that in the
higher organism, growth is a highly complex process, and that in
consequence it cannot be represented mathematically. It is be-
cause of the very fact that growth is a complex process that
mathematical analysis of the experimental data is necessary.
Corollary to this must be the recognition of the fact that since
growth is not a simple process, no one mathematical formula will
be adequate for full description2 and no one method adequate for
complete analysis.

Our purpose in the present note is to illustrate on a series of
data collected by one of us (1919) the application of inter-periodic
correlation coefficients to certain phases of the problem of growth.

Before passing to the analysis, which is the special purpose of
this paper, definition of the terms which will be used and a note

1 Citations of literature may be traced from Reed's paper.

2 Those who consider the possible adequacy of a single equation take the
ground that if it be possible to represent the growth of an organism by a
simple equation, it may be by virtue of the fact that during growth the various
(often conflicting external) factors which affect the living substance are inte-
grated by the organism.

243

244

J. ARTHUR HARRIS AND H. S. REED.

on the nature of the data on which the statistical methods are
illustrated are in order.

By growth stage we mean any given moment of time at which
a series of organisms are measured. It is, therefore, synonymous
with age during the growth period. The absolute size of the
organism or of one or more of its parts at a given growth stage
is the only character of the organism available for consideration.

By growth period we understand the period of time elapsing
between the sth and the s -f- nth growth stage.

The increase in size during any such period we shall designate
as a growth increment.

By relative growth increment, /",-«, we understand the ratio of
the growth increment, /, to the absolute size of the individual at
stage, r, where r and 5 are any two successive stages.

Turning now to the question of the original data as given in
Table I. of Reed's (1919) publication we note from a study of the
physical constants for absolute size in Tables I. and II. that there
is an increase in the mean height of the plants up to the 77th day.

TABLE I.

STATISTICAL CONSTANTS FOR SIZE AT VARIOUS GROWTH STAGES.

Growth Stage.

Mean.

Standard Deviation.

Coefficient of Variation.

7

I7.Q7I

I 617

Q O

14 .

76.728

4 786

17 •»

21 ...

67.845;

8 Q72

172

28

O7 672

IA 677

I ? O

•JS .

I7O 724

TO 17/1

Td 7

42

168.707

24.801

14 7

49

20=;. 107

72 76o

16 o

56. .

22Q 672

77 842

16 ^

63.

2A7 1AZ

A2 Z.7A. *

17 2

70. .

2ZI 776

A-i /177

T 7 7

77

2s^.8lO

47 767

17 2

The increase from the 63d to the /oth and from the 7oth to the
77th day is relatively slight, being only 443 cm. or 1.79 per cent,
of the height for the 63d day in the first case and only 2.03 cm.
or 0.81 per cent, of the value for the 7oth day in the second case.
The difference between the 84th day and the 7/th day is negli-
gible. In view of the fact that there is no appreciable growth in

INTER-PERIODIC CORRELATION.

245

the sense in which the term is used here between the 77th and the
84th day, this period will be left entirely out of account in the
calculation of the correlations for the following discussions.

Furthermore by considering the constants for growth incre-
ments as shown in Table II., we note that the coefficients of varia-

TABLE II.

STATISTICAL CONSTANTS FOR GROWTH INCREMENTS FOR VARIOUS GROWTH

PERIODS.

Growth Period.

Mean Increment.

Standard Deviation. Coefficient of Variation.

7 to 14

18.397

3-704

20.5

14 tO 21

7I.CI7

5.164

16.4

21 to 28

20.827

7.907

26.5

28 to "is

•33. 052

7.505

22.7

^ to 42 . ,

37.983

11.578

30-5

A2 tO AO

^6.690

14.266

38-9

J.O tO ^6

24.276

16.540

68.1

s6 to 6^? .

17.672

13.803

78.1

6^? to 7O .

4-4^1

4-713

106.4

70 to 77

2.0^4

5.096

250.5

tion for growth increments from the 63d to the 77th day are
abnormally great. This may be in part due to biological causes,
but it is doubtless due to a considerable extent to the relatively
large error of measurement when the increment is very small in
comparison with the size of the organism. If this be true, we
should expect the correlations for actual size for the 63d to the
84th day to be about the same as those for the immediately preced-
ing growth stages, but the correlations for growth increments may
be expected to be of little value.

The problems which may be considered will be presented and
discussed seriatim.

II. ANALYSIS OF DATA.

PROBLEM I. The correlation between the absolute size of the
organisms at its several periods of development.

When examined at an early stage of development, organisms
are found to differ among themselves in size. The same is found
to be the case when the same series is measured at a later growth
stage or at maturity.

In the biological analysis of the phenomenon of growth a prob-

246 J. ARTHUR HARRIS AND H. S. REED.

lem of great importance is that of the causes which bring about
the differences in size observable at any stage of development, or
after growth has entirely ceased. Are individuals which are
found to be small at maturity those which were small initially and
have remained so from the beginning, or may the growth rate of
an individual change during the course of its development to such
an extent that it may vary its position in the series under investi-
gation from time to time? That the latter is to some extent the
case we know from general observations on human children. The
problem to be solved is that of the quantitative magnitude of the
relationship between the size of the individual at different stages
of development.

The nature of the biological problems to be investigated has
been stated in earlier work, and an attempt has been made to
solve them by grouping plants according to quintile (Pearl and
Surface, 1915) or quartile (Reed, 1919) position in the culture
to which they belong and ascertaining the quartile or quintile in
which they fall at different stages of growth.

This method has the disadvantage that all the individuals,
whatever their size, are lumped together in four or five groups.
In this method of treatment, small differences between two indi-
viduals are, therefore, given as much significance as large ones,
providing they are large enough to throw the two individuals into
different quartiles or quintiles.

An alternative method, which will completely obviate this dif-
ficulty, is to determine the correlation between the sizes of the
individual at different periods of growth. The possible correla-
tions between the absolute size of the individuals in the n dif-
ferent stages of growth of the Hcliantluis plants are shown in
Table III.

The coefficients in this table can be best understood by first ex-
amining those for the relationships between the sizes of the plants
near the period of maturity, and then passing to the relationships
between the sizes of the plants at earlier stages.

Considering first of all the coefficients in the lower right-hand
corner of the table, we note that all the coefficients are very high,
denoting practically perfect correlation. This is the relationship
which would be expected for a period when the organism has

INTER-PERIODIC CORRELATION.

247

w
j

a
<

H

tn

M

u

<

EH

t/2

K
H
£

O
«

O

H

H

H
<

Pn

tLH
O

w

o

5

E

u

K

H

H

n

_;

a

K
Bi
O

U

a

rt

t/5

oo

CO
O

r-

OO
O

oj

OO
O

o

OO
O

O M
t- O

OO

O
O

ON

O
O

oo

O

oo

O

oo

O

O
O

-H

oo

o

oo

o

oo

o

-HS.-H

O

o
-H

00 00 00 !>• ^O ^f H

ooooooo

.
O\iOMOOOro O

"

t^ Tf- 10 ro oo

OOOOt^O^t
OOOOO

-H S.-H5 ^ ° -H 5^

oooo

•H^g -H 3-H S-H

.

OS J> 00 ON

o

IJ
H

oO

O

oo

JJCNJJ
"tl. ,A "H

'°MI^

OOO

O\
M
O

OOOOOO

-H .-H-H o(i\ S -H S -H

o

_nO\

'Tj-

o

xC

o

r-

oo

oo

oo

O

o

o

O M M

ooo

d

0^ -H8 -HM

oo

oo

O\

oo
t^

O

vo
oo

O

r^
oo

O

oo
oo

O

oo
oo

O

oo
oo

O

oo
oo

O

rt
55

(*5

>O

248

J. ARTHUR HARRIS AND H. S. REED.

practically attained its adult size and in which there is relatively
little change from one week to another.

As we follow the correlations between the later periods and
preceding periods back, we note that there is a regular decrease
in the values of the correlation coefficients. This may be best
shown by summarizing the results graphically in diagram I.

In the graph the correlation of the size of the organism at each

l-o

•3
•8

14- Zl 28 35 42. 49

AGE OF PLAA/T (STAGE)

DIAGRAM i.

56 63

70

growth stage with its size at every antecedent growth stage (shown
at the bottom of the diagram) is shown on the scale of correlation
at the left by points marking the magnitude of the correlations
for each of the growth stages. The pitch of the lines connecting
the points for the i_j.th to the 77th growth stage shows the rapid
decrease in the magnitude of the correlations as the stages be-
come more widely separated in time.

INTER-PERIODIC CORRELATION.

249

The same type of diagram may be used to show the relation-
ship between the size at early and at later growth stages. Dia-
gram 2 shows the distribution of the magnitudes of the correla-
tions for sizes of the individuals at the 7th to the 7oth day (stage)
and the size at subsequent growth stages.

From these lines it is clear that the correlations between size

1-0
•9
•8

r

14 21 28 35 42. 49 56
/JG£ OF PLAA/T (STAGE)

DIAGRAM 2.

63

70

77

at antecedent and subsequent periods decrease as the periods be-
come more widely separated in time. This is true without ex-
ception for every period which furnishes evidence upon the
question.

The coefficients are, however, positive in sign throughout, thus
suggesting (though in some cases not proving) that throughout
its growth period the size of the plant bears some relation to its
size when first measured. This result is in agreement with the

25O J. ARTHUR HARRIS AND H. S. REED.

findings of Webber (1920) in regard to the growth of Citrus
stock.

PROBLEM 2. The correlation between the growth increments
of the organism during the several growth periods.

Our second problem is to determine whether there is a corre-
lation in growth increments as well as in actual size of the or-
ganism. We shall thus answer the question whether the organism
which grows more rapidly than the average during one growth
period will grow more rapidly than the average in other growth
periods and whether the organism which lags behind the average
in its rate of growth during one growth period will also lag be-
hind during other growth periods.

Little has heretofore been done towards the statistical treat-
ment of growth increments. This is probably in part due to the
arithmetical difficulties of computing the constants for incre-
ments, but if the moments and product moments be taken about
zero as origin in computing the coefficients required under Prob-
lem i above, the calculations for growth increments are easily
made by the use of formulae given elsewhere (Harris, 1920).

The symmetrical table showing the relationship between the
actual growth increments for all of the combinations of growth
periods appears as Table IV. This table shows positive and sta-
tistically significant correlation coefficients for closely associated
periods throughout the season up to and including the period for
the 63d to the /oth day. The coefficients for the period from
the 7Oth to the 7/th day cannot in general be considered statis-
tically significant in comparison with their probable errors.

Examining these results in a little greater detail, we note that
the nine coefficients showing the relationship between the growth
increments of successive weeks (the constants bordering the diag-
onal cell of the symmetrical table of constants) are all positive in
sign and with the exception of the last (showing the relationship
between the growth of the period from the 63d to /oth and that
between the /oth to 7/th day) all are statistically significant. The
eight coefficients measuring the correlations between the growth
increments of weekly periods which are separated by one week
are also without exception positive, but are lower in magnitude
and less certainly statistically significant. For periods more

INTER-PERIODIC CORRELATION.

251

o

o
t^.

r-» oo
co oo

o q

r f

C\

CO

q

r

O CO

cc oo

q q

r r

00
CO

o

+

ON 10

oo oo

q q

-i- +

OO
CO

o
4i c

4'

t

e.

0

en

q q
f l'

o

-H c

r

q q

r r

o
^ 0-

r

o q

JJ ^r -U ""
Tl TJ- Tl C

+ +

t

\

>

CO

co

o

41 o\

o
41

PH PERIODS.

m

0

oo oo
q q

II ^ 11 ^

Tl O> Tl tn

M 6 ° H

00 O

O M

+ +

00

o

£*

M

f

oo co

q q

4l PO 41 r-
J PO (vj t— o

rs o

r +

oo

o

o ^

10

+

oo

O

N
+

o

o

-i-

oo

0

< ij O
S Tl (N

+

EVERAL GROW

10

2

oo oo

o o

-rl^-H H

&\ t-H

O O

r +

OO

o

1 _n oi

r

CO 00

o q

r +

t^~

o

O iy

+

oo
q

4i t

5 §! '

+

r-

o

M'

41

oo

0

?j{?

o

+.

in

w

pH

H

O

z

1— 1

p

Q

Period.

o

OO 00

0 0

jj t~ jj r

N u ^* f

M O

r r

oo

q

*" _U '

3 Tl

r

CO ro

q q
+ +

0

1
4

t— oo

O O

41 41

co

q

2 41

f

oo

o

3 41 rr

2 OO o

+

1 — 1

w

M

H

INCREMENTS i

Growth

o
tn

oo oo

o q

r +

oo

q

+

O

0

0 jj r-
+

ro

q
41

00 00

o q
' 0 ^ r-

+ +

oo

o

0 _M

f

oo

q

r

THE GROWTH

IO
0
CO

OG OO

q q
+ +

10

o

•T; -O

+

0

q
+

oo

0

~- jj c

•O Tl I

° 5 '
41

oo oo

o q

f l'

o
o

ro jj
fO Tl

r

oo

2-fl S
r^. o\ _;

f

)NS BETWEEN

oo

o

M

w

oo *o

q q

+ +

XI

10 oo

q q

JJ I-J _M V

+ +

oo

o

D JJ

r

oo oo

q q

r f

oo

q

r

oo
q

r

H

M

O

PO
ir>

O

M CO
OO 00

O 0

ON
oo

O

O\ 00
co oo

0 O

oo
oo

0

CO

oo

O

OS
Hi

0

U

o

H

is

-H

°°- 41 M 41

ii

t- jj ro II
c^i Tl w T

o^°§

:•«

PO II ro
O\ Tl 10

4-

+

4- 4-

i

4- +

i

1

1O

o

PO
co
o

O\ I—
oo oo

o o

oo
oo

0

CO 00

oo oo

0 0

oo
oo
O

oo

O

4-

-H

?4t

« 41 2"4I

?)41

CT4I o 41

<>41

£-H S;

£

M £

n z ° z

" 2

M ?Mi

0 &

° PO M

4-

+

+ 1

1

1 4-

1

1

o *^-

U U

Tt M
M N

0 0

CO

o

IO M

0 0

o

O PC

0 0

o

0

o

*• 2"

H

oo 10

O)

o o

ro

0

252 J. ARTHUR HARRIS AND H. S. REED.

widely separated in time the correlations are in part positive and
in part negative in sign.

Thus from the results as a whole it appears that the incre-
ments of successive periods are generally positive and fairly
highly correlated when the periods show actual growth incre-
ments. Thus the zone of coefficients lying along the diagonal
cell are positive and generally fairly high. When the periods are
separated by any considerable length of time the coefficients are
generally insignificant in magnitude and may, as a matter of fact,
be either positive or negative in sign.

The relationship may be brought out by determining the aver-
ages of the correlation coefficients, with regard to sign, for the
increments of periods separated by various lengths of time. The
results are as follows.

Period of Separation Number of Correlations Average

(Weeks). Averaged. Correlation.

0 9 + -5009

1 8 + .2240

2 7 —.0334

3 6 —.1236

4 5 -.1640

5 4 —.1033

6 3 —.0077

7 2 —.0585

8 i — .1360

If we disregard the cases in which there are less than five co-
efficients to be averaged, we note a steady decrease in the magni-
tude of the correlation coefficient. Periods of growth which are
successive or separated by only one week have positively corre-
lated growth increments. Periods which are more widely separ-
ated show negative correlations of the increments.

The relationship between the coefficients in Table IV. may be
clarified by diagram 3 which shows the relationship between four
of the ten growth increments and each of the other ten incre-
ments. The increments selected as a " first variable " in the cor-
relation are the first, fourth, seventh, and tenth. This has the
advantage of representing the first and the last growth incre-
ment, and of leaving undrawn no more than two successive incre-
ments. The figures are aligned according to the ten increments
representing the " second variable " of the correlation.

INTER-PERIODIC CORRELATION.

253

- +-60 K

- +-*o £

|

PER/OTJ 7 TO /4

-+•20 K
- g

00 *

BO

i < < i i i _i

•60

•40

•20

•GO'

•20

•40

k PER 101 28 JO 35

o

t-H
P

+ •40

-PER/OH 49 70 SS

•+•20

•00

•-•20

1

PfRlOU 70 TO 77

•00

--•20 ,

ft:
I

•STAGE

7 14 21 28 35 42 49 56 63 70 77

254 J- ARTHUR HARRIS AND H. S. REED.

The graphs for the first, fourth and seventh increment show
clearly the shift in the position of the maximum positive corre-
lation from the earlier to the later periods as the " first variable "
is chosen from the later periods. The same is shown less clearly
by the correlations for the tenth increment, but there the coeffi-
cients are very small, presumably because growth has practically
ceased.

It is clear, therefore, that plants which are growing more
rapidly during any period of development will 'grow more rapidly
during a closely associated subsequent period of development but
that there is little or no relationship, or even a negative relation-
ship, between the rate of growth of the organisms studied at con-
siderably separated periods of time.

Since the correlations for absolute growth increments are so
small for all except successive periods of time, it seems unneces-
sary to deal at present with the relative growth increments, i.e.,
with the growth increments expressed as a fraction of the size of
the organisms at the beginning of the growth period.

PROBLEM 3. The correlation between the absolute sise of the
organism at given stages of development and subsequent growth
increments.

In the higher plant organism rate of growth at any period
must be supposed to depend to some extent upon plastic ma-
terials synthesized by the more nearly mature portions of the
same individual. Thus one might expect to find a relationship
between the actual size of the organism at any stage of growth
and the rate at which the organism increases in size during a sub-
sequent period.

We have determined the possible correlations between the ab-
solute size of the organism at different periods and the growth
increment of the organism during subsequent growth periods.
The coefficients are presented in Table V. This shows positive
correlation between the actual size of the organism at every stage
of development from the 7th to the 7Oth day and the increase in
the size of the organism during the following week. The mag-
nitude of the correlation is of the order ?- = o.45 to r==o.6o for
the Jth, I4th, 2ist, and 28th day. For these growth stages the
correlation between actual size and the subsequent growth incre-

INTER-PERIODIC CORRELATION.

255

W
J
H
<

H

Q
O

H

O

OS

o

H
<

U)

H
Z
H

S

W

H

O
2

to

a

H

O

«

O

W

H
H

a
o

H
M

O

tn
2

O

H

6

oo t-

cooo t^r^-oooo o\O\

oo co

oooooooooooooooo

•

o o

OOOOOOOO

^

O

_u 1" JJ

~n M ~n ^

Oil MTloOTi •^TIO~H OTlooTl O~H O

*••

O . ' i_i

' *"O ' to .— 1 Tf .J 10 ,_j ro k«4 *"* o ^ o ^" o

Q

O ^3"

O r~^ cs ^f ON f^* O O

*"•

M 1-1

wO^i—iOOOO

f f

i' r r r r \ \

oo oo

OO O O\ oo co Ox f^*

00 00

oo oo t^- r^- oo GO oo

o

q q

o o o o o q o

t-s

_M ^J* _U

O IJ W H ^" jj I>» jj O\ ij f^" 1 I tN i 1 CO

o

"71 N Tl

F*- Tl o\ Ti M TJ H "Ti ro Tl M Tl ro Tl Tf

M X C^

Q IH X ro pj O\ri~f*r5^'^ fO^O ^M

CO

o "o

OO CO O) ^^ MO W W

O M ro f) d O M

f \

r r r r i' r +

oo oo

OO 00

o q

oo oo oo oo co oo

oooooo

o

it t~- _i|

Tl rf Tl

S+la-riS'-H 5-H^-Ho-H M

MO

^t ° to

'-'oo'0' ?°' c?"^0' S5°'co ^

tr>

O 0

o o M o q H

1 +

+ 1 1 1 + +

oo oo

oo r— ^f t^ oo

^o

oo oo
0 O

oo oo oo oo oo

0 '0 0 0 O

o

-H Jo-H

g'-H^-fi ^.-H S-H^-HW

-*-1

\O « O

MNr>csMr^^toMONd

o

•* " 0\

TJ- N M ro t^

"*

q q

o H p> H q

f \

i' r f r +

oo oo

oo oo ON t>

oo oo

OO OO OO f^

•

0 O

o o q q

o

-U M 44
Tl 'T Tl

sOjj^H^ll1^^1^

O "Tl t^ Tl ON™ OTlsO

•o

N

^ ° ON

Mvo>c'ooo'o"o''Or*'

_o

O 0

O O O ro

1-1
01

—

1 1

111 +

J3

>o t^~

oo oo ro

%

00 00

oo oo oo

2

•

o q

O O O

O

0

jj co _n
Tl oo Tl

O Tl ^f Tl t^ Tl O

IT)

O M ^"

M TJ- \O lO

CO

M M

O O M

f \

f + +

O\ ON

r- IN

00 OO

oo r—

ID

q q

q q

CO

11 C^ II

M _M r^- ij oo

o

Tl ro Tl

9 ^r J? 9

oo

W O

° u~i M ro

M

O 0

I-H TT

\ +

+ +

•* M

W">

oo oo

0

co

q q

q

o

Tl oo "H

O jj O
to Tl ON

M

ro OO

ro M t>-

M <N

+

M ^O

oo ^

•

q q

N

O

M

"H 10 Tl

^

H

o

0

o

-ri °

10

SEJS

t- n

M oo vo d ^^ ^O fO O

256 J. ARTHUR HARRIS AND H. S. REED.

ment is clearly significant in comparison with its probable error.
The coefficients are lower for the 35th and the 42d day, but are
probably statistically trustworthy. Beyond this period there
seems to be no relationship between the size of the organism and
its growth rate in an immediately following period.

For the first two stages of growth measured, the 7th and the
I4th day, there may be a significant correlation of the order
r=:-{-. 285 between size and growrth increments during the sec-
ond following week.

The coefficient of correlation between size and the increment
in the second week following is also positive for the 2ist and
28th day, but neither of these values may be considered statis-
tically significant in comparison with its probable error. Finally
for the first stage (seventh day) there may be a significant cor-
relation between absolute size and growth increments during the
third week following (r- +.239 for increment for 2ist to 28th
day). Other than this the coefficients are for the most part sta-
tistically insignificant in comparison with their probable error.

Summarizing the preceding statements as a basis for further
analysis, we note that for the first six growth stages (7th to 42d
day) there is a significant positive correlation between the size of
the organism at the given stage and the growth increment of the
following week. For the first two growth stages (and possibly
in the third where r/Er= =1.77) there is a significant correlation
between the size of the organism and the growth increment in
the second subsequent week. Finally, for the first stage only,
there is a significant positive correlation between size and growth
increment in the third subsequent week.

Disregarding these 9 coefficients and the 4 positive but not sig-
nificant Correlations between the sizes at the several growth stages
and the growth increments of the following week, we may note
the following facts concerning the remaining 42 coefficients.

Of these 42 coefficients 36 are negative while only 6 are posi-
tive in sign. Of the 6 positive coefficients only that between
actual size on the 2ist day and growth increment between the
28th and 35th day (already considered above) is as large as its
probable error. Of the 36 negative coefficients 18 are larger than

INTER-PERIODIC CORRELATION. 257

their probable error, and 5 of these are over twice as large as
their probable error.

There is, therefore, clear evidence that the subsequent growth
of the higher plant organism is measurably conditioned by its
size. In general the larger individuals grow more rapidly in im-
mediately subsequent periods, but somewhat more slowly than
the average in more distant subsequent periods.

While a detailed discussion of the relation of these results to
the theory that growth may be satisfactorily described by the
curve of an autocatylic reaction falls quite outside the scope of
this paper, it must be noted that negative correlations between
actual size at a given stage and the growth increments of certain
subsequent growth periods might be expected. As Reed and
Holland (1919) have pointed out the plants attained about half
their final height at about the thirty-fourth day. From this time
on the increments were decreasing. Plants which had attained
more than the average size at this period would, therefore, of
necessity, make smaller average increase in size in later periods.

The number of individuals measured is not sufficiently large to
carry the analysis farther.

III. RECAPITULATION.

The purpose of this paper has been to illustrate on the basis of
a specific series of data the value of the inter-periodic correlation
coefficient in the analysis of the phenomena of growth.

The analysis shows that in the case of a series of Heliantlins
plants the actual size of the individual at any stage of develop-
ment is closely correlated with its size at other closely associated
stages of development. The magnitude of the correlation rapidly
diminishes as the growth stages become more widely separated in
time. Thus the final size of the organism is but to a slight extent
determined by its initial size.

The correlation between successive growth increments is posi-
tive in sign and statistically significant, with the general average
of ^=.501. The correlation for increments of weekly periods
separated by one week is on the average only about ~r = = .225.
For periods more widely separated than this the correlation be-
tween growth increments is on the average negative in sign.

258 J. ARTHUR HARRIS AND H. S. REED.

Thus plants which are growing more rapidly during one period
of development will grow more rapidly during a closely asso-
ciated period, but there is little or no relationship between the
growth increments of more widely separated periods.

The growth increment of the organism is positively corre-
lated with its size at an immediately preceding stage. In the
early stages of growth, the growth increments of two or even
three subsequent periods are positively correlated with the initial
size of the organism.

LITERATURE CITED.
Harris, J. Arthur.

'20 Formula for the Calculation of the Correlations of Size and of Growth
Increments in the Developing Organism. Proc. Soc. Exp. Biol. and
Med. In press.
Pearl, R., and Surface, F. M.

'15 Growth and Variation in Maize. Zeitschr. Indukt. Abst. u. Verer-

bungslehre, 14 : 97-203.
Reed, H. S.

'19 Growth and variability in Heliantluts. Amer. Jour. Bot., 252-271.

'19 The Growth Rate of an Annual Plant, Heliantlnts. Proc. Nat. Acid.

Sci., 5: 135-144.
Reed, H. S.

'20 The Xature of the Growth Rate. Jour. Gen. Physiol., 2: 545-561.
Webber, H. J.

'20 Selection of Stocks in Citrus propagation. Bull. Univ. Calif. Coll. Agr.,

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, MARCH l6, IC>2I.

THE CHROMOSOMES OF PSEUDOCOCCUS NIP^E.

FRANZ SCHRADER,
BRYN MAWR COLLEGE.

INTRODUCTION.

In the course of some work on sex determination in the dif-
ferent species of Pseudococcus — a genus of the Homoptera—
very peculiar conditions were met with in the chromosome be-
havior. These peculiarities were observed especially in Pseudo-
coccus nipcc (identified by H. Morrison, Bureau of Entomology),
and it is of this species that the present account is given. A more
detailed report, covering other species of Pseudococcus as well,
is reserved for a later paper.

Most of the material was fixed in Allen's modification of
Benin's fluid. On the whole, fixation is more or less difficult ;
and at best the cells are somewhat small. The main features are
clear cut, however, and hardly to be mistaken. I am indebted to
Professor E. B. Wilson and to Professor C. E. McClung who
examined some of my slides and offered helpful criticism.

THE CHROMOSOMES IN THE FEMALE.

The number of somatic chromosomes in the female is ten, with
little or no size and form differentiation. Counts are made with
little difficulty in various cells, but oogonial cells furnish of
course the best criterion. Generally the chromosomes are counted
most easily just before they .have become arranged in a meta-
phase plate. There can be no doubt as to their number (Fig. 6).
A detailed study of the maturation phenomena in the female was
not made. Suffice it to say that five tetrads are formed and that
these are normal in appearance ; they are very much like those
formed in the oogenesis of many other Homoptera (Fig. /).
The reduction process is thus probably not unusual.

SPERM ATOGENESIS.

The somatic number of chromosomes in the male is also ten,
and as in the female these seem to be alike in size and shape.

259

26O FRANZ SCHRADER.

Such chromosome counts were generally made in cells of the de-
veloping nervous tissue where division phases are common (Figs.
3 and 4).

Spermatogonial divisions seem to be completed with compara-
tive speed, for specimens which show them are not plentiful.
Just as in the somatic cells the number here is undoubtedly ten

(Fig. 5).

The stage following the spermatogonia seems to be much longer
in duration. The cells increase perceptibly in size during this
time. The earliest phase observable shows some flocculent masses
of lightly staining chromatin irregularly distributed through the
nucleus. At one point, always at the periphery of the nucleus,
there is a more deeply staining mass. Nothing concerning the
structure of this can be made out and its shape is variable (Fig.
8). With progressive development this deeply staining mass
undergoes a few, very definite, changes. In successive steps it
appears that a number of more or less irregular lumps is evolved.
Still massed at first, these gradually become separated and then
it is certain that they are five in number (Figs. 8 to n). It is at
this latter stage that a split is occasionally visible in some of them,
but with increasing condensation this again becomes obliterated.
Throughout this development these five bodies retain a definite
tendency to remain in close proximity to each other, and this
tendency is one that persists also through subsequent stages.

In the meanwhile the flocculent and more lightly staining chro-
matin has also undergone development. Before the denser mass
has become evolved into five distinct bodies, this chromatin has
been transformed into a fine network of threads. Apparently
these are polarized toward the dense mass (Fig. 9). Like the
leptotene threads of other forms, these threads shorten and
thicken, a process accompanied by a progressive increase of their
staining intensity. Polarization is finally lost, and already at this
stage it becomes apparent that the number of shortened threads
is less than ten (Fig. 10). As the threads continue their process
of shortening, they are counted with greater ease, and in such a
stage as shown in Fig. 1 1 it becomes certain that they are five in
number. Like the denser bodies, these sometimes show a lon-
gitudinal split.

THE CHROMOSOMES OF PSEUDOCOCCUS NIPyE. 261

Both the denser bodies as well as the threads continue pro-
gressive condensation, and the former reach their final form some
time ahead of the latter (Fig. 12). They are then somewhat
oblong in shape, and take the haematoxylin stain with great inten-
sity and there can be no doubt of their chromosomal nature.
Somewhat later the erstwhile threads have also assumed this
form, and there are then ten of these bodies or chromosomes,
identical in size and shape. Those first evolved continue to be-
tray a certain affinity for each other, and in the metaphase plate
constitute a central group around which the other five chromo-
somes become ranged in no definite order (Figs. 13 and 14).
Aside from this very characteristic grouping, the only difference
between the two sets of chromosomes that is apparent consists in
the rate at which they evolve or the stage which is the starting
point of their development.

Throughout this development, there has been no trace of a
tetrad formation. The general features of the case indicate that
the split which was spoken of as occurring at one stage is nothing
more than preparation for the equation division or else something
of the nature of the " Querkerbe " observable in lower Crustacea.

Division now occurs in ordinary manner and ten chromosomes
go to each pole (Figs. 15 to 17). The arrangement of chromo-
somes in the daughter cells is not absolutely certain, although fig.
1 6 indicates that there also the characteristic grouping is retained.
Figures like these are too rare to admit of any definite conclusion
however. At any rate, the time in which such an arrangement
persists must be very short, for the chromosomes are generally
found in a more or less irregular heap (Fig. 18).

The division just described is undoubtedly equational in char-
acter. Following it there seems to be no intervening further
development in the chromosomes of the daughter cells. Instead,
they begin to scatter in a longitudinal direction. This process is
not entirely irregular however for it results in their separation
into two groups of five each (Figs. 1 8 to 22). It is a remarkable
feature that these two groups are each characterized by a distinct
and different arrangement of their component chromosomes.
The group going to one pole assumes the form of a V or a

262 FRANZ SCHRADER.

triangle, while the sister group which goes to the opposite pole is
circular or lumped in arrangement (Figs. 21 and 22). This
grouping is so constant and has been observed in so many speci-
mens, that no mistake seems possible, and the conclusion seems
inescapable that it is of some significance. The telophase of this
anomalous division still shows traces of the arrangement, but
these are soon lost as the chromosomes of each daughter cell dis-
tribute themselves around the periphery of the nucleus. Their
number here is undoubtedly five (Fig. 24). This initiates the
formation of the spermatids in which the chromosomes gradually
loose their staining reaction. No study of the subsequent stages
was made except to determine that there is no sign of degenerat-
ing or abortive cells nor a size dimorphism in the spermatozoa.

SOMATIC CHROMOSOMES.

Returning to the somatic tissues, it may be remarked here that
although the number of chromosomes in each sex is the same,
their arrangement differs in the two sexes. This is especially
noticeable in the developing nerve tract, where in the male the
cells in the resting stage show a relatively large nucleolus like
structure (Fig. 2). This is not to be seen in the same tissue in the
female where cells show only the flocculent chromatin peculiar to
that phase (Fig. i). That the nucleolus-like structure in the
male nerve cells is nothing but the group of five chromosomes
mentioned in the description of the spermatogenesis becomes
almost certain in metaphase plates found in the same tissue.
Figs. 3 and 4 show such grouping without a doubt.

Exactly the same feature is observable in spermatogonial plates,
though the size of these renders them less favorable (Fig. 5). In
contradistinction, oogonial plates have no such arrangement, and
even in such a late stage before division as shown in Fig. 6 the
chromosomes are arranged in no definite order.

DISCUSSION.

An interpretation of these observations is perhaps not out of
place. It is here given with the idea that it should not affect the
observations however it may be received and is advanced in a
speculative way.

THE CHROMOSOMES OF PSEUDOCOCCUS NIP.E. 263

As has already been stated, oogenesis very probably follows
ordinary lines. The ten chromosomes constituting the diploid
number are composed of five homologous pairs, and these synapse
and form tetrads. Reduction is very probably normal, and re-
sults in a pronucleus with five chromosomes.

In the spermatogenesis, the spermatogonial divisions, like the
somatic divisions, also occur in orthodox manner. This is ap-
parently not true of the meiotic divisions however. In explana-
tion of these, the best hypothesis is one which views the various
developments in the light of sex chromosomal behavior and is as
follows :

The central group of chromosomes which appears in the growth
stages of the male as the more densely staining mass contains sex
chromatin, equally distributed among the five chromosomes. The
remaining chromosomes, which stain lightly at first, represent
what may be regarded as purely autosomal chromatin. Granting
this, and the assumption does not appear unjust in the light of
what has been described, the seemingly peculiar development
becomes a natural consequence. Just as in the spermatogenesis of
the various Orthoptera and Hemiptera, the sex chromosomes
always stain more or less intensely, and as far as observable do
not go through the various stages of thread formation. That
such formation may occur earlier, or in a restricted sense even
while the dense mass is still irregular in outline, is not ruled out
by any means. The autosomal chromatin on the other hand goes
through all the usual steps, culminating in the formation of five
chromosomes. The sex chromatin contained in the five grouped
chromosomes will tend to explain their grouped arrangement,
since again as in the Hemiptera, multiple X or X and Y chromo-
somes show a tendency to remain in close proximity during
development.

If now the sex chromosomes in Pseudo coccus nipcc are re-
garded homologous in every way to the autosomes, except that
each carries a certain amount of sex chromatin, the subsequent
behavior is just as would be expected. I may mention here that
such sex chromosomes would imply a more intimate union of sex
and autosomal chromatin than is illustrated bv such a case as

264

FRANZ SCHRADER.

Menniria (McClung, '05) where the sex chromosome and the
autosomes are distinct, but the former is attached to one of the
latter. It is at present unnecessary to go into the relation of the
two conditions, though very possibly they represent two distinct
steps in the phylogeny of sex chromosomes.

Their subsequent behavior is more or less analogous to that of
the X Y pair in other forms. This pair does not form a tetrad
in the ordinary sense simply because its members are not homo-
logous, or better perhaps, because neither has a true synaptic
mate. When as in the homozygous state both members of a pair
of sex chromosomes are homologous, synapsis and tetrad forma-
tion occur just as in the autosomes. This fact is plainly borne
out in the oogenesis of many Hemiptera (Morrill, '10) as well as
in the growth and maturation phenomena of the eggs of Pscudo-
coccus nipcc. It is thus to be assumed that if in the present case
of the spermatogenesis of Pseudococcus nipce the sex chromatin
were distributed equally over the ten autosomes, the pairs would
be homologous and tetrads would be formed in the usual way.

The cytological evidence indicates nothing that should render an
equation division exceptional in nature, and it does indeed occur in
the usual manner. The second division witnesses reduction in
that the autosomes carrying sex chromatin go to one pole while
the purely autosomal chromosomes go to the opposite pole.
Taking recourse to a parallel case once more, attention may be
drawn to the two X chromosomes in Syroniastcs which always go
the same pole in reduction (Wilson, '09). Similarly, the multiple
X of the Reduviidae always goes to one pole, although this is not
an exactly parallel case since it is probably the product of frag-
mentation of a single X.

Thus to repeat what has already been intimated for the present
case, the distribution of the chromosomes to their respective poles
in the reduction division may be explained on the ground that we
are concerned with five pairs of chromosomes. The members
of each pair are homologous except for the fact that one of them
in each instance carries a certain amount of sex chromatin. The
presence of the latter does not influence the behavior of the
chromosome pairs in reduction and the members of each pair go

THE CHROMOSOMES OF PSEUDOCOCCUS NIPJE. 265

to opposite poles. Its presence does however prevent haphazard
distribution in that the five chromosomes carrying this sex chro-
mation tend to remain clustered or grouped and therefore go to
the same pole.

Although more or less contrary to the cytological evidence
furnished by other groups of insects, it may not be amiss to sug-
gest the possibility that in animals with haploid males each
chromosome carries a certain amount of sex chromatin. It
follows that the diploid female would then represent 2 X, whereas
the haploid male would represent I X. In the haploid male the
reduction division is not truly abortional as has been supposed,
but is merely a division in which these sex chromatin carrying
chromosomes go to one pole while the opposite pole receives no
chromosomes simply because the mates to these chromosomes are
absent. It is of interest to note that the straggling or lagging so
often observed in the sex chromosomes of various insects is
paralleled by the scattered and irregular distribution of the chro-
mosomes on the spindle of this division in the Hymenoptera. And
lastly, such irregular distribution is found also in the reduction
division of Pseudococciis niptr.

Pseudococciis nipcc thus would stand half way between forms
with haploid males where every chromosome carries sex chro-
matin, and forms in which the sex chromatin is carried in very
few chromosomes and there is little numerical variation in the
chromosomes of the two sexes. In other words, half of the
chromosomes in the males carry sex chromatin.

Although superficially an instance of Weismann's postulated
ideal type of reduction in which the diploid number of chromo-
somes is halved without previous syndesis, the spermatogenesis
of Pseudococciis nipcc nevertheless follows the commonly ac-
cepted lines of meiosis. The apparently exceptional behavior
can be explained as due to an extreme mode of sex chromatin
distribution and is not a unique example of the Prim&rtypus of
reduction. It may be remembered that Goldschmidt ('05 and
'08) gave this name to an instance of Weismann's simple type
which he thought to have discovered in Zodgonus minis. The
Schreiners ('08) examining Goldschmidts slides believed to have

266 FRANZ SCHRADER.

found a serious error in his counts of somatic chromosomes,
which they believed in reality to be 24 and not 10 as he had
reported. Furthermore, reduction occurred in the ordinary way,
just as in Tomoptcris. Gregoire ('09) in going over the same
slides maintained that the Schreiners were correct in that the
case was one of ordinary reduction, but that they in turn had
made an error in the chromosome counts. The somatic number
is about 12, and the reduced number 6. Lastly Wassermann
('n, '12, and '13) procured new material and concluded that
Gregoire's counts had been correct. He did not agree with
Gregoire as to the mode of synapsis however, and apparently was
unable to reach a final conclusion in this regard himself. Al-
though he thus does not believe that the question has received a
definite settlement, the fact remains that Zoogonus does not
represent the simple type of reduction that Weismann advanced
in a hypothetical way.

If my hypothesis is correct, the male of Pscudococcus nip a; is
heterozygous in that it has five sex-chromatin carrying chromo-
somes and five chromosomes purely autosomal in character.
Crossing over would not occur in these chromosomes. It would
occur however in the female, in which the ten chromosomes are
composed of five homologous pairs. If the male represents i X,
the female with ten sex-chromatin carrying chromosomes repre-
sents 2 X.

SUMMARY.

1. The diploid number of chromosomes in Psendococcits nipce
is ten in both sexes.

2. In the maturation of the egg, five tetrads are formed and
reduction is probably normal.

3. In the spermatogenesis, five chromosomes are developed
before the others, and 'these tend to remain grouped together.

4. No tetrads are formed, and in reduction five chromosomes
go to one pole (supposedly those evolved first) and five to the
other.

5. Explanation of this seemingly anomalous behavior is to be
sought in the fact that five of the chromosomes carry sex
chromatin.

THE CHROMOSOMES OF PSEUDOCOCCUS NIP.E. 267

6. The case is not so much to be regarded as an illustration of
Weismann's ideal type of reduction, as an exceptional example
of reduction due to unusual sex chromatin distribution.

BIBLIOGRAPHY.
Goldschmidt, Richard.

'05 Eireifung, Befruchtung, und Embryonalentwicklung des Zodgonus

minis. Zool. Jahrb., Bd. 21.
'08 Die Chromatinreifung der Geschlechtszellen des Zodgonus mints und

der Primartypus der Reduktion. Arch. f. Zellf., Bd. 2.
Gregoire, Victor.

'09 Le reduction dans Ie Zodgonus mints Lss. et le " Primartypus." La

Cellule, Tome 25.
McClung, C. E.

'05 The Chromosome Complex of Orthopteran Spermatocytes. BIOL. BULL.,

Vol. 9.
Morrill, Charles V.

'10 The Chromosomes in the Oogenesis, Fertilization, and Cleavage of

Coreid Hemiptera. BIOL. BULL., 19.
Schreiner, A., und K. E.

'08 Neue Studien, V. Die Reifung der Geschlechtszellen von Zodgonus

mints. Vidensk. Selks. Schr. I. Math.-Naturv. Kl.
Wassermann, F.

'n Ueber die Eireifung bei Zodgonus mints Lss. Sitz.-Ber. d. Ges. f.

Morph. und Phys., Mtinchen.
'12 Zur Eireifung von Zodgonus mirus, ein Beitrag zur Synapsisfrage.

Ver. d. Anat. Ges. Bd. 26.
'13 Die Oogenese des Zodgonus mints Lss. Arch. mikr. Anat., Bd. 83,

Abt. 2.
Wilson, E. B.

'09 Studies on Chromosomes. IV. Jour. Exp. Zool., 6.

All drawings made with a Spencer 15 eyepiece and Zeiss 1.5 objective with
the exception of Figs, i and 2, where a 10 eyepiece and 2 mm. objective were
used.

268 FRANZ SCHRADER.

PLATE I.

1. Cells in nervous tissue of the female.

2. Cells in nervous tissue of the male.

3 and 4. Metaphase plates from nerve tissue of the male.

5. Spermatogonial plates.

6. Oogonial cell.

7. Tetrads prior to polar body formation in the egg.
8 to 12. Growth stages prior to first division.

13. Metaphase plate of first division.

BIOLOGICAL BULLETIN VOL. XL.

PLATE I.

1

v

•;rr?.

f-f

'-.'I

. - •
*'£%&*'.

•/ /
• , *

'•

/

' /

7

-

10

V

11

12

13

FRANZ SCHRADER.

27O FRANZ SCHRADER.

PLATE II.

14. Metaphase plate of first division.

15. Side view of plate.

16 to 18. First division (equational).
19 to 23. Second or reduction disvision
24 and 25. Spermatids.

BIOLOGICAL BULLETIN, VOL. XL.

PIATE II.

14

'

15

16

17

18

19

20

'..

21

22

,* ;N
'I,

0

23

*

'

24

** *

.

V -^ 4

v ^ /

25

FRANZ SCHRADER.

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, MARCH l6, IQ2I.

OBSERVATIONS ON THE LARV^ OF CORETHRA
PUNCTIPENNIS SAY.

CHANCEY JUDAY.i

INTRODUCTION.

As part of a general problem relating to the biological produc-
tivity of lakes, a quantitative survey of the bottom fauna in the
deeper portions of Lake Mendota at Madison, Wisconsin, was
made between the early part of May, 1916, and the middle of
August, 1918. This survey included only the macroscopic forms,
such as the insect larvae, the Oligochseta, and the Mollusca. The
investigation showed that the full-grown larvae of Corethra punc-
tipcnnis Say constitute the principal element of the bottom popu-
lation in the daytime during the greater part of the year. For at
least three quarters of the year, in fact, they not only far outnum-
ber all of the other forms combined, but they also exceed them in
total weight. The great abundance of these larvae thus makes
them a very important factor in the biological complex of the
lake.

Samples of the bottom were obtained by means of a modified
form of the Ekman dredge ; the opening of this instrument cov-
ered an area of 473 square centimeters. The mud obtained in
each haul of the dredge was washed through a gauze net having
meshes fine enough to retain all of the macroscopic forms. The
material secured by the net was transferred to a jar and was then
taken to the laboratory where the various organisms were sorted
out and enumerated in the living state. The average live and
dry weights of the various forms were also obtained, as well as
the percentage of ash.

Observations were made at five regular stations located in
water having a depth of 20.5 meters to 23.5 meters. These sta-
tions were widely separated in order to secure a fair average of
the density of the bottom population in the deeper portions of the

1 Notes from the Biological Laboratory of the Wisconsin Geological and
Natural History Survey No. XXI.

2JI

2J2

CHANCEY JUDAY.

lake. The results obtained at only one of these stations, desig-
nated as Station II., are considered here, however, because the
other four were not visited regularly during the winter months.
From April to June inclusive, the number of larvae found at Sta-
tion II. was from 10 per cent, to 20 per cent, larger than the gen-
eral average of the five stations during these months, but in
August the general average was larger than the numbers at Sta-
tion II. In the six sets of averages obtained for the last three
months of the year in 1916 and 1917, three of the averages for
Station II. were larger than the corresponding averages of the
five stations and three were smaller. Thus, the numbers ob-
tained at Station II. were somewhat larger than the general aver-
age for the deeper part of the lake as a whole during the first
half of the year, but they were somewhat smaller from July to
September and substantially the same from October to De-
cember.

NUMERICAL RESULTS.

The average number of Corethra larva? per square meter of
bottom is shown for the different months of the year in Table I.
Large numbers of these larvae live over winter ; in fact they are

TABLE I.

THE NUMBER OF CORETHRA LARVAE PER SQUARE METER OBTAINED FROM THE
MUD ix THE DAYTIME AT STATION II. DURING THE DIFFERENT MONTHS

OF THE YEAR.

In all months except January and February the numbers represent averages
of two to nine samples.

Year.

January.

February.

March.

April.

May.

June.

ioi6. .

17,300

4,??O

IOI7. .

2S.34O

2^,74.0

30,380

33,230

I7,9OO

l6,OOO

1918

23, 44O

20,160

l8,O4O

18,170

0.43O

Year.

July-

August.

September.

October.

November.

December.

1016. .

3,380

7,400

17,600

26,900

26,90O

IQI7 . .

6,080

8OO

-2,740

II,5OO

I7,70O

24,8OO

1018. .

2,820

060

more numerous from November to April than at any other time
of the year. During this interval the numbers range from ap-
proximately 18,000 to 30,000 individuals per square meter. At

OBSERVATIONS ON LARVAE OF CORETHRA PUNCTIPENNIS. 273

this time of the year there is no loss from pupation and the losses
from other causes are not great enough to reduce the number of
these larvae very materially, so that the number remains uni-
formly high during this period of time.

The ice usually disappears from the lake during the first week
in April and soon after this event the larvae begin to pupate. As
the temperature of the water rises, the rate of pupation increases
so that an appreciable decrease in the number of larva is noted
for the month of May. With the further advance of the season,
pupation becomes still more common and this results in a very
marked decrease in the number of Corcthra larvae in late June as
in 1916 and in 1918, or in early July as in 1917. This decline in
numbers continues until the minimum of the year is reached in
August, more especially during the first half of this month. A
minimum of 295 larvae per square meter was noted at Station II.
on August 2, 1917, while the average in March of this year was
a little more than one hundred times as large. (See Table I.)

Small swarms *of adults appear in May and in early June, but
the great flights are correlated in time with waves of very active
pupation in late June, in July, and in early August. Thus, enor-
mous swarms appear from time to time during the latter period.

During late August and especially in September there is a
slackening in the process of pupation and correlated with this is
an increase in the number of larvae. The increase is most marked
during the second half of September and in early October, but
the numbers do not reach the maximum point until November or
December. The largest number of larvae obtained in any of the
samples was 33-8oo individuals per square meter on December
21, 1917.

Pupae were not noted in the samples of mud until about the
middle of June, or at the beginning of the more active period of
pupation ; thereafter they appeared regularly until late August.
A maximum of 2,890 pupae per square meter was found on June
28, 1917, while the second in rank was 1,370 individuals per
square meter on July 9, 1917.

274 CHANCEY JUDAY.

DEPOSITION AND DEVELOPMENT OF EGGS.

According to Muttkowski the adults emerge at night, "begin-
ning early in the evening and continuing through the night. In
the morning, if the lake is quiet, the females can be seen on the
surface, ovipositing through the surface film." Eggs deposited
by females kept in insect cages sink to the bottom of the aquarium
and experimental evidence indicates that those deposited at the
surface of the lake also sink to the bottom. Mud from Station
II. where the water is 23.5 meters deep and also from Station I.,
located in water 18.5 meters deep, was washed through a sieve
with meshes fine enough to remove all of the larvae ; this sifted
mud wTas then placed in aquaria. At the end of five days a dozen
small Corcthra larvae had appeared in the material from Station
II., while five small larvae were noted in the other bottom ma-
terial at the end of a week, thus showing that the mud from both
stations contained eggs.

Eggs that were deposited by females kept in captivity hatched
within forty-eight hours when the temperature of the laboratory
ranged from 21° to 24° C. The temperature of the lower water
in the deeper portions of the lake is much lower than this, how-
ever, and the eggs which reach the bottom in these areas probably
do not develop so rapidly. The bottom temperature at Station
II. in summer, for example, ranges from slightly more than 9°
in some years to about 14° in other years. On the other hand,
eggs deposited in water not exceeding five meters in depth are
subject to temperatures of 20° to 25° in July and August so that
they probably hatch about as promptly as those kept under lab-
oratory conditions.

Another factor that may retard development in the deeper
water is the absence of free oxygen. Usually all of the dissolved
oxygen below a depth of 18 meters is used up by the middle of
July, after which no oxygen is available in this region until Oc-
tober. Thus, all of the Corethra eggs which reach the bottom in
water that is 18 meters deep or more during this period must de-
velop under anaerobic conditions if they develop at all. This
anaerobic stage covers the greater part of the most active repro-
ductive period of this insect and approximately 30 per cent, of

OBSERVATIONS ON LARWE OF CORETHRA PUNCTIPENNIS. 275

the area of the lake lies within the 18 meter contour. For two
or three weeks in August, in fact, a little more than half of the
lake bottom is subject to anaerobic conditions. No attempt was
made to ascertain the effect of this lack of oxygen on the devel-
opment of the Corethra eggs. Many cocoons of the Oligochaet
Limnodrilus wrere noted in the bottom material during this period,
however, and the eggs in them seemed to be developing normally
in the absence of oxygen. This fact suggests that the eggs of
Corethra may also develop normally under anaerobic conditions.
The young Corethra larvae were not noted in the series of net
catches until the last week in June, though the eggs of the first
adults each year undoubtedly hatch at an earlier date than this.
They were found regularly in the net catches from the latter part
of June to the first week in October.

BEHAVIOR.

In 1917 and 1918 net hauls were made regularly at three of
the stations before the mud catch was taken in order to see if
any of the full-grown larvae occupied the water in the daytime,
but the results were entirely negative. Some of these hauls were
made as early as 8 : 30 A.M. and others as late as 4:30 P.M., so
that these observations covered the chief portion of the day. It
was found also that the full-grown larvae deserted the water on
cloudy days as well as on clear days.

A series of observations made at Station II. during the after-
noon and evening of July 16, 1917, showed that the full-grown
Corethra larvae had not emerged from the mud by 7 : 30 P.M., or
just about sunset. At 8:00 P.M., or half an hour after sunset,
133 larvae and 88 pupae per square meter of lake surface were
found in the water. By 8:30 P.M., or one hour after sunset,
these numbers had increased to 3,945 larvae and 442 pupae. At
the latter hour the full-grown larvse had reached the surface of
the lake, thus showing a vertical migration of 23.5 meters during
an interval of about one hour.

A similar set of observations was made in 1920 beginning at
5 :45 P.M. on June 10 and continuing until 5 :3O A.M. on June
ii. On the former date sunset came at 7:36 P.M., standard
time, and sunrise on the following day at 4:18 A.M. No Co-

276 CHANCEY JUDAY.

rethra larvae were obtained from the water on June 10 until 7:15
P.M., at which time a catch yielded 22 individuals per square
meter of lake surface. Fifteen minutes later, or just a few min-
utes before sunset, this number had risen to 176 larvae per square
meter, and by 7:50 P.M., or about a quarter of an hour after
sunset, the number was 1,576. At 7:36 P.M. the larvae had not
invaded the upper 10 meters of water, but they had ascended to
the 10-15 nieter stratum. They did not appear at the surface
until about an hour and a quarter after sunset, so that the rate of
upward migration was somewhat slower than that noted in 1917.
Pupae reached the surface at 9:00 P.M., or approximately an
hour and a half after sunset. The largest number of both larvae
and pupae found in the water during this set of observations was
noted in a catch taken at 10:00 P.M.; of the former there were
4,730 individuals per square meter of lake surface and of the
latter 287. By 11:00 P.M. the numbers had declined to 2,100
and no respectively; the numbers were substantially the same as
these at 2:00 A.M. on June 11.

Larvae were still found in the upper meter of water at 3 : 30
A.M., but they had disappeared from the upper 10 meters by
3:47 A.M. and only one individual was obtained in a catch taken
from the 0-15 meter stratum at 3:50 A.M. Practically, then,
they deserted the upper 15 meters of water during a period of
about 20 minutes. It should be noted, also, that this downward
migration was not due to direct sunlight since it took place at
least half an hour before sunrise. The larvae were still occupy-
ing the 15-23 meter stratum in considerable numbers, since a
catch at 3:55 A.M. yielded 1,658 individuals per square meter
in that region ; the same catch contained 88 pupae per square
meter also. The number of larvae in the lower water then gradu-
ally diminished, the last disappearing between 4 : 45 and 5 : °°
A.M. According to these results, then, the full-grown Corcthra
larvae enter the bottom mud by the end of the first half hour after
sunrise and they remain there until about sunset.

Samples of mud taken at 6:00 and at 7:00 P.M. on June 10,
1920, yielded an average of 2,720 larvae and 55 pupae per square
meter of bottom ; as a result of the migration into the water these
numbers had declined to 1,400 larvae and 22 pupae per square

OBSERVATIONS ON LARWE OF CORETHRA PUNCTIPENNIS. 2J7

meter at 8:00 P.M., while the samples obtained during the next
three hours yielded from 1,600 to 2,100 larvae. The latter num-
ber was also found at 3:00 A.M., bvit it rose to 2,665 at 4:00
A.M. and to slightly more than 3,000 per square meter at 5:00
A.M. Thus the mud contained from one half to two thirds as
many larvae at night as were found there in the daytime.

For a certain period after they hatch out, the behavior of the
young larvae is very different in the daytime from that of the full-
grown individuals ; that is, the former occupy the lower water
during the daylight hours instead of the mud, being found in the
lower part of the mesolimnion and in the hypolimnion. The
young larvae migrate into the upper water at night just as the
full-grown ones do. It has not been definitely determined just
how long this difference in behavior lasts ; only rarely was an
individual found in the mud which was estimated to be only one
third as large as a full-grown larva and frequently individuals
were obtained from the water wrhich were recorded as half
grown. Thus, it appears that the young larvae inhabit the lower
water in the daytime instead of the mud until they are approxi-
mately one third grown, or perhaps a little larger. Muttkowski
states that the larval period lasts from six to seven weeks in the
summer broods; on this basis it may be estimated that the dif-
ference in behavior between the young and full-grown larvae con-
tinues for the first ten days or two weeks of the larval period.

A series of catches wras made with a plankton trap on August
2, 1917, for the purpose of ascertaining the vertical distribution
of the small larvae. The results are shown in Table II. No larvae
were found in the upper 8 meters, but they appeared at 10 meters
and at greater depths. The maximum number, 489 individuals
per cubic meter of water, was obtained at a depth of 18 meters,
which was about the middle of the hypolimnion. Somewhat
more than 88 per cent, of the total number of individuals occu-
pied the 15-20 meter stratum.

Some results obtained on Devils Lake, Wisconsin, show that
the behavior of the full-grown larvae of Corcthra plumicornis
Fabricius1 differs in the daytime from that of C. punctipennis in

i Dr. ]. R. Malloch kindly identified this larva.

278 CHANCEY JUDAY.

TABLE II.

THE NUMBER OF YOUNG CORETHRA LARVAE PER CUBIC METER OF WATER AT
DIFFERENT DEPTHS OF LAKE MENDOTA ON AUGUST 2, 1917.

Those obtained at 10 meters and 12 meters were recorded as very small
and for the other depths the individuals were estimated to be from a quarter
to a third as large as full grown larvae.

Depth in Temperature, Number of Larvae

.Meters. Degrees C. per Cubic Meter.

8 19-8 O

10 17-4 44

I2 l6.0 67

15 14-5 20°

18 13-6 489

20 13-5 3ii

23 13-3 22

Lake Mendota. In the former lake two net catches on May 25,
1917, which were made in the deepest water, namely, 14 meters,
gave an average of 422 full-grown C. plumicornis larvae per
square meter of surface, while two hauls of mud at the same
place yielded an average of 433 individuals per square meter.
That is, these larvae were substantially equally divided between
the water and the mud at about 10: oo A.M. on a bright morning
when the water was so transparent that a white disc 10 centi-
meters in diameter did not disappear from viewr until it reached
a depth of 8.6 meters. In other words, the day distribution of
the larvae of C. plumicornis was practically the same in Devils
Lake as the nocturnal distribution of the larvae of C. punctipennis
in Lake Mendota.

While the larvae of Corethra punctipcnms give a prompt nega-
tive reaction to light, yet it hardly seems probable that their ex-
tensive depth migration in Lake Mendota, even including a
descent into the mud, is a simple light phenomenon. The trans-
parency of the water is usually low in summer; a white disc 10
centimeters in diameter generally disappears from view at a depth
of two meters to about four meters at this season of the year,
which indicates that the light is cut off rather rapidly by the upper
strata of water. On the morning of June n, 1920, for example,
the disc reading was 4.25 meters. A pyrlimnimeter has been used
to determine the rate at which the sun's energy is cut off by the
upper strata of the lake. The results obtained with this instru-

OBSERVATIONS ON LARVJE OF CORETHRA PUNCTIPENNIS. 2J9

ment indicate that the intensity of the illumination at a depth of
23 meters on a clear day, between n : oo A.M. and I : oo P.M., is
about equal to that produced by full moonlight at the 'surf ace of
the lake. During the early forenoon and the late afternoon, as
well as on cloudy days, the illumination is much smaller than this.
For some time before sunset, the bottom stratum must be sub-
stantially in total darkness, yet the observations show that the
emergence of the larvae from the mud is very closely correlated
in time with the setting of the sun.

Not only does the illumination in the bottom water become
very small in the late afternoon, but there is a further protection
from light afforded by the bottom ooze in which the larvae remain
concealed during the day. To what depth the larvse penetrate
the loose mud is not known, but in the laboratory they readily
burrow down to a depth of a centimeter or more. The dim light
which reaches the bottom in the deeper portions of the lake can
penetrate the ooze only to a very slight extent at most, even dur-
ing the brightest part of the day, and this raises the very interest-
ing question as to what stimulus causes the larvae and pupae to
emerge from the mud so promptly and regularly about the time
of sunset. No definite data bearing on this point have yet been
obtained.

These larvae are eaten with avidity by many fishes and their
habit of occupying the mud in the daytime may thus serve a very
important purpose from the standpoint of protection from such
enemies. A further protection is afforded by the disappearance
of the dissolved oxygen in the hypolimnion. Usually by the first
of August very little free oxygen remains in this stratum, which
makes the lower water unfit for the permanent occupation of the
larger forms which prey upon these larvae. In spite of the lack
of oxygen, however, P'earse and Achtenberg found that the yel-
low perch — Perca flavesccns (Mitchill) — invades the lower
water and feeds upon these larvae. While these fish survive for
a period of two hours in water that contains no dissolved oxygen,
these authors state that it is doubtful whether a perch is able to
feed for more than a few minutes at a time under such conditions.
It seems probable, therefore, that the Corethra larvae are not
eaten as freely as they might be if anaerobic conditions did not

280

CHANCEY JUDAY.

prevail in the lower strata of the deeper water. Also, the ab-
sence of dissolved oxygen in the hypolimnion serves as a protec-
tion to the young larvae which occupy this region in the daytime
for a certain period after they hatch out.

NUMBER IN SHALLOWER WATER.

The larvae of Corcthra pnnctipennis show a decided preference
for the deepest portion of Lake Mendota. In the daytime, they
are much more abundant in the mud where the water reaches a
depth of 20 meters or more than they are in the shallower areas.
It was found that the average number of larvae within the area
bounded by the 20 meter contour line was more than three times
as large as the average for the region lying between the 8 meter
and the 20 meter contours, while the number obtained in areas
where the water did not exceed five meters in depth was prac-
tically negligible.

Some three hundred samples were taken in series which ex-
tended from the shallow water to the deep water ; that is, from
a depth of 8 meters or 10 meters down to a depth of 20 meters.
The results of four sets of these observations are shown in Table
III. It will be noted that there was a marked increase in the

TABLE III.

THE NUMBER OF CORETHRA LARVAE PER SQUARE METER OF BOTTOM AT DIFFER-
ENT DEPTHS IN FOUR SETS OF OBSERVATIONS WHICH WERE MADE IN 1917.

Date.

Depth in
Meters.

Number.

Date.

Depth in
Meters.

Number.

May 15

10. S

820

September 24 . .

IO

110

June 22 . ....

12.5
15-5

18

2O
IO

1, 6OO
4.640
4,810
7,800
IOO

October 24

12

15
18
20
12

250
5-500
7.490
13.380
85

12
IS
18.5
20.5

85
i, 080
i, 710
3,610

IS

18

20

2,740
11,650
15.930

number of larvae correlated with the increase in the depth of the
water. On May 15, for example, the sample taken at 15 meters
yielded about six times as many as the one at 10 meters, while

OBSERVATIONS ON LARV.E OF CORETHRA PUNCTIPENNIS. 28 1

that at 20 meters gave nearly ten times as many as the latter. On
September 24, the differences were fiftyfold and more than a
hundredfold, respectively, and on October 24 the number was
nearly two hundred times as large at 20 meters as at 12 meters.

Just how these larvae are able to constantly maintain such a
marked difference in numbers in favor of the deep water is a
puzzling question. Their method of locomotion would not lead
one to expect them to travel very far of their own accord should
they reach the shallow areas, yet it seems probable that many of
them are carried into the shallow water by the currents when
they migrate into the upper strata at night. This would be true
especially on windy nights. Table III. shows that a very small
number of larvae is found at 10 meters as compared with the
deep water and 39 per cent, of the area of the lake lies outside the
10 meter contour line. The number is usually not much larger
at 12 meters than at 10 meters and the former divides the area of
the lake approximately into halves. The outer or shallower half
of the lake, then, is very sparsely populated by these larvae, but
it is not clear just how the number is kept so small in comparison
with the inner or deeper half of the lake.

In a large proportion of the former area the bottom does not
consist of material in which the larvae can readily conceal them-
selves in the daytime, being composed of sand, gravel, and rock,
so that the tendency would be to avoid these areas. On the other
hand, the larvae are no more abundant in the shallow portions of
protected bays where a muddy bottom suitable for concealment is
found at a depth of only 5 meters or 6 meters. The difference
can scarcely be attributed to a proportionally unequal distribution
of eggs between the two regions because very large numbers of
egg bearing females are found over the shallow water as well as
along the shore ; it seems probable, therefore, that enormous num-
bers of eggs are deposited in the shallow areas.

When the larvae migrate into the upper strata of the lake at
night, the direct currents tend to carry them into the shallow
water, but the return currents, on the other hand, will aid more or
less in bringing them back to the deep water. Apparently the
chief factors governing the distribution of the Corcthni larvae

282 CHANCEY JUDAY.

between the shallow half and the deep half of the lake are (i) an
active migration, (2) the currents — direct currents on the wind-
ward side of the lake and return currents on the leeward side,
(3) a relatively greater loss in the shallow water due to pre-
datory enemies.

NUMBER IN OTHER LAKES.

For purposes of comparison similar quantitative studies of the
bottom population were made regularly in Lake Monona and in
Lake Waubesa during 1917. The former is only one kilometer
from Lake Mendota and has nearly as great a maximum depth,
namely, 22.5 meters. Lake Waubesa lies about seven kilometers
southeast of Lake Mendota, but it is a much shallower body of
water, having a maximum depth of only a little more than u
meters. In Lake Monona only about one tenth as many Corethra
larvae were found as at corresponding depths and times in Lake
Mendota ; in some instances the difference was more than a hun-
dredfold in favor of the latter lake. In the deepest part of Lake
Waubesa the number varied from about the same as that at 1 1
meters in Lake Mendota to only a third or a quarter as many ; but
the deeper water of Lake Mendota yielded from forty to a hun-
dred times as many larve as the deepest portion of Lake Waubesa.
Bottom material has been obtained from about a dozen other Wis-
consin lakes and in all of them the Corethra population has been
relatively small, which seems to indicate that Lake Mendota offers
a particularly favorable habitat for these larva?.

GRAVIMETRIC RESULTS.

Between June, 1916, and April, 1917, more than fourteen
thousand larvae of Corethra punctipennis were picked out of the
material collected in Lake Mendota and they were dried for the
purpose of making a chemical analysis of them. The average
amount of dry matter per individual for this number was 0.251
milligram. The average weight was also determined for eleven
other lots of larvae containing from 100 to 300 individuals each.
These averages ranged from a maximum of 0.311 milligram per
larva, dry matter, in June to a minimum of 0.182 milligram in
early September. The results of these weighings are shown in

OBSERVATIONS ON LARV.E OF CORETHRA PUNCTIPENNIS. 283

Table IV. The higher average of dry matter in the June material
may be due to a larger proportion of chitin in the larvae just
before they pupate. In August and early September the average
size of the larvae seems to be smaller than that of the winter brood
and this is confirmed by the weights. At the height of the
pupating season the summer larvae pupate when they are distinctly
smaller than the individuals which live over winter. The live
weights of the smaller lots were also determined, as shown in
Table IV., and they indicate that about 91 per cent, of the living
animal consists of water.

The live weight of the June pupa? was only about n per cent,
larger than that of the June larvae, but the dry weight of the
former was nearly twice as large as that of the latter. (See
Table IV.) This marked difference in the dry weight was prob-
ably due to the presence of a larger amount of chitin in the pupa.

TABLE IV.

THE AVERAGE WEIGHT OF A SINGLE INDIVIDUAL OF CORETHRA PUNCTIPENNIS IN
MILLIGRAMS, TOGETHER WITH THE PERCENTAGES OF WATER AND OF ASH.

Form.

Month.

Live Weight in
Milligrams.

Dry Weight in
Milligrams/

Per Cent, of
Water.

Per Cent, of

Ash.

Larva ....

February. . .
May

3-06
1.1O

0.250
0.264

91.72
Q2.I2

7.06

7 O6

June

I.I e;

O.1II

80. I*?

7 11

September. .
October. . . .
November. .

2-57
2.83

3-20

0.182
0.264
0.285

92-93
9O.66
91.03

9-53
8.62
8.10

Pupa. .

June . . .

1. $2

O ?74

8l 71

C go

Adult

June . . .

O.7=C

O.427

41 12

c go

The adults yielded a much smaller live weight than either the
larvae or the pupae because they possessed a much smaller propor-
tion of water. Their dry weight was greater than that of the
larvae but smaller than that of the pupae. The adults used for
this weight were obtained from a large swarm on June 29, 1918,
when pupation was very active, but there was no means of ascer-
taining their age ; their weight probably decreases somewhat with
age, and they live for a period of three to five days.

The ash of the larvae varied from a minimum of about 7 per

284 CHANCEY JUDAY.

cent, of the dry weight to a maximum of 9.5 per cent., the former
being noted in February and the latter in September. The pupae
and adults yielded substantially the same percentages of ash, but
these percentages were much smaller than in the larvae.

There is more or less overlapping of the summer broods, which
makes it difficult to estimate the number of larvae produced during
this season, but the winter crop of larvae may be estimated with
some degree of accuracy. In this investigation twenty-two
samples of mud were obtained at the five regular stations in deep
water during the month of November and they yielded an average
of 17,350 Corcthra larvae per square meter of bottom. Five
samples were secured in December also, and they gave an average
of 21,900 individuals per square meter; but four of these samples
were taken at Station II. which usually gave a larger yield than
the other four stations. According to these figures, the early
winter population of Corcthra larvae within the 20 meter contour
may be conservatively estimated at 18,000 individuals per square
meter. Between October and May the live weight averaged 3.1
milligrams per larva and the dry weight 0.266 milligram. Apply-
ing these weights to the above population gives a live weight of
55.8 grams per square meter, or 558 kilograms per hectare, which
is equivalent to 497 pounds per acre, and a dry weight of 4.8
grams per square meter, or 48 kilograms per hectare, equivalent
to 42.7 pounds per acre.

Muttkowski states that there may be two generations of
summer larvse in addition to the winter generation ; but, since the
former average somewhat smaller in size than the latter, the total
weight of the summer broods is probably not greatly in excess of
that of the winter brood. That is, a live weight of 1,200 kilo-
grams per hectare (1,070 pounds per acre) would be a conserva-
tive estimate for the total annual production of Coretlira larvae
in the deeper part of Lake Mendota ; on this basis the dry weight
would amount to somewhat more than 100 kilograms per hectare,
or approximately 90 pounds per acre. These figures apply only
to that portion of the lake which lies within the 20 meter contour
line, since the larvae are found in very much smaller numbers
in the shallower water.

OBSERVATIONS ON LARVAE OF CORETHRA PUNCTIPENNIS. 285

The 20 meter contour encloses an area of 664 hectares which
would give an annual crop of larvse amounting to substantially
797 metric tons, live weight, for this portion of the lake, or a dry
weight of about 67 metric tons. The live weight of all other
macroscopic inhabitants of this area was 92.3 metric tons and
their dry weight was about 19 metric tons.

As previously indicated, the population of Corethra larvse in
the region between 8 meters and 20 meters averaged about one
third as large per unit area as that in the deeper water. In order
not to overestimate the annual production of the shallower water,
the area lying between the shoreline and a depth of 10 meters
may be omitted from the calculation since the number of larvae
found in this region is small ; in addition, also, the average
between the 10 meter and 20 meter contours, comprising an
area of 1,738 hectares, may be reckoned as one quarter instead of
one third as large as that of the deep water. On this basis the
live weight becomes 300 kilograms per hectare and the dry
weight 25 kilograms, thus making the annual crop of Corethra
larvae in this portion of the lake a little more than 521 metric
tons, live weight, or about 43 metric tons of dry material. These
results combined with those obtained for the deep water area give
a total annual production of 1,318 metric tons of living larvse
which would yield no metric tons of dry material.

CHEMICAL RESULTS.

The results of the chemical analysis of the larvae are shown in
Table V. and they are stated in percentages of the dry weight.

TABLE V.

RESULTS OF THE CHEMICAL ANALYSIS OF THE LARV.E OF CORETHRA PUNCTI-
PENNIS STATED IN PERCENTAGES OF THE DRY WEIGHT.

Nitrogen.

Crude Protein
(NX6.25.)

Ether Extract

(Fat).

Crude Fiber
(Chitin)

Per Cent.
Ash,

10.74

67.12

9-45

6.15

7.96

The percentage of nitrogen is notably high, which means a cor-
respondingly large proportion of crude protein. The percentage

286 CHANCEY JUDAY.

given in the table does not include the nitrogen in the crude fiber
(chitin) which amounted to 0.46 per cent. In comparison with
this the larvae of Chironomus tentans yielded a much smaller
percentage of nitrogen, namely, 7.36 per cent.

The larvae yielded a fairly large amount of fat (ether extract),
namely 9.45 per cent, of the dry sample. Together the crude
protein and the fat constituted more than 76 per cent, of the dry
material. From the standpoint of quality, this large proportion
of these two. excellent food materials gives the larva of Corcthra
punctipennis a very high rank as a source of food for other

organisms.

LITERATURE.
Birge, E. A., and Juday, C.

'n The Inland Lakes of Wisconsin. The dissolved gases of the water and
their biological significance. Wisconsin Geological and Natural His-
tory Survey, Bulletin No. XXII, pp. xx •-•- 259. Madison.
Juday, C.

'04 The Diurnal Movement of Plankton Crustacea. Trans. Wis. Acad.
Science, Arts and Let., Vol. XIV., Part 2, pp. 534-568. Madison.

'08 Some Aquatic Invertebrates that Live under Anaerobic Conditions.
Trans. Wis. Acad. Science, Arts and Let., Vol. XVI., Part i, pp. 10-16.
Madison.
Muttkowski, R. A.

'18 The Fauna of Lake Mendota. Trans. Wis. Acad. Science, Arts and

Let., Vol. XIX., Part i, pp. 374-482. Madison.
Pearse, A. S., and Achtenberg, Henrietta.

'20 Habits of Yellow Perch in Wisconsin Lakes. Bulletin of the Bureau
of Fisheries, Vol. XXVI. , pp. 295-366. Washington.

Vol. XL. • June, IQ2I. No. 6

BIOLOGICAL BULLETIN

MORPHOLOGY AND ORIENTATION OF THE
OTOCYSTS OF GONIONEMUS.1

L. J. THOMAS.

INTRODUCTION.

Gonionemus is used as an example of the Hydro-medusa in
many zoological laboratories of this country. In spite of this
fact relatively' little is known regarding the morphology and
orientation of the sense organs of the representatives of this
genus. Both the location and the structure of the otocysts have
been very imperfectly treated in the literature. Frequently
morphological and experimental treatises mentioning the otocysts
refer the reader to the works upon other genera, the otocysts of
which are presumed to be essentially similar to those of Gonio-
nemus. In studying specimens of Gonionemus vcrtcns Ag. and
of G. Jintrbachii Mayer the writer was impressed by the lack of
agreement between his direct observations and the published
statements by various early workers. These discrepancies led
the writer to investigate the problem further in an attempt to
discover the source of the statements which have been so gen-
erally incorporated into the literature and to determine, if pos-
sible, the precise structure and relationships of the otocysts in
this genus.

On the Pacific coast of this country G. vcrtens A. Ag. occurs i'~>
the Puget Sound region and on the Atlantic coast G. uiurbachii
Mayer is found in abundance in the Eel Pond at Woods Hole,
Mass. Many references in the older literature incorrectly refer
to the Atlantic species as G. zfcrtcns because at that time the
Atlantic form was not considered as specifically distinct from the

i Contributions from the Zoological Laboratory of the University of Illinois,
No. 181.

287

288 L. J. THOMAS.

type of the genus. In 1895 Murbach mentioned certain differ-
ences between the two forms and in consideration for his work
upon the Atlantic species A. G. Mayer (1901) described it as a
distinct species under the name of Gonionetnus murbachii.

The terms otocyst, statocyst, and lithocyst have been used
interchangeably throughout the literature on Ccelenterates. In
this paper the term otocyst is used in referring to the entire struc-
ture which is supposed to function as an organ of equilibrium.
A careful study of the otocysts in Gonionemus has revealed the
fact that the terminology ordinarily employed in describing the
otocyst is entirely inadequate for an intelligible description of
the parts and understanding of their functional relations. In
Gonionemus the otocyst is not a simple vesicle enclosing an
otolith as has usually been considered the case. From a morpho-
logical point of view there is considerable evidence that in Gonio-
nemus the large vesicular structure is merely an adaptation for
the protection of the true sensory apparatus all of which is
lodged within the structure that has ordinarily been termed the
otolith. This obvious confusion of terms renders a detailed de-
scription of the organ necessary.

Details of the organization of the otocyst are shown in Text
Figure I. The wall of the primary vesicle ('/>?•) is the outermost
wall of the entire organ and encloses a fluid-filled cavity (c)
within which the spheroid (s) is suspended by the primary
pedicel (pf>~). The thick wall of the spheroid encloses the fluid-
filled cavity which is designated as the secondary vesicle (si'}.
Within the secondary vesicle rests the otolith (o), free to move
about within the fluid-filled chamber. Extending from the distal
end of the primary pedicel to the membranous lining of the
secondary vesicle is a distinctly differentiated region to which the
term secondary pedicel (sp) has been applied.

This investigation was carried on under the general direction
of Dr. H. J. Van Cleave, to whom the writer is greatly indebted
for suggestions and for securing the material and the identifica-
tion of the species under consideration. Individuals of Gonio-
nemus murbachii were obtained through the Supply Department
of the Woods Hole Marine Biological Station for comparison

MORPHOLOGY OF OTOCYSTS OF GONIONEMUS.

289

with those of G. vcrtcns upon which the greater part of the work
was done. Specimens of G. vertens collected at Friday Harbor,
Washington, were submitted to Dr. Alfred G. Mayer and to Dr.

TEXT FIGURE i. General organization of the otocyst of Gonionemus: pv,
primary vesicle; c, fluid filled cavity; s, spheroid; pp, primary pedicel; si>,

secondary vesicle; o, otolith ; s p, secondary pedicel.

/

Henry B. Bigelow, both of whom very kindly verified the tenta-
tive identification of the species.

While the present paper deals primarily with the finer struc-
ture of the otocyst the latter part is devoted to a discussion of
the orientation of the otocyst in the organism.

METHODS OF STUDY.

Specimens preserved in formalin were too opaque for accurate
observations of the otocysts. Serial sections and toto-mounts
of the bell margin, including ring canal, otocysts, and tentacles
were prepared for the detailed study of the morphology and the
orientation of the otocysts. The toto-mounts were made by

2QO L. J. THOMAS.

trimming the tentacles close to the bell margin with fine scissors
and then clipping this rim, with the velum included, free from the
remainder of the bell. Such preparations were stained with
borax carmine and mounted in damar. The otocysts, though
embedded in the mesoglea, were by this method sharply defined
and readily visible for accurate observations. In some speci-
mens prepared in this manner the finer details of structure could
be determined. By careful manipulation, mounts prepared in
this manner showed practically no distortion. Individuals 9 to
15 mm. in diameter were dehydrated, cleared, and mounted in
damar with a shrinkage of i mm. or less in diameter. This
shrinkage was obviously sustained uniformly by the various
tissues for no evidences of wrinkling or distortion of parts were
observable in the finished preparations. All drawings were made
with the camera lucida from prepared mounts and sections.

STRUCTURE OF THE OTOCYST.

Murbach (1903: 205) dismisses the anatomy of the otocysts
in Gonioncinus with the footnote: "The finer anatomy is natur-
ally omitted. The description of the nervous system and the
otocyst given in ' Das Nervensystem mid die Sinnesorgane der
Medusen,' O. u. R. Hertwig, text, pp. 48-69, Plates 4 and 5, fits
very nearly those of Gonioneinus." A careful examination of
the Hertwig descriptions and drawings of otocysts in other
genera reveals many points where a closer " fit " might be desired
if the descriptions and drawings are to cover the conditions found
in Gonioncinns.

Superficial examination under a compound microscope reveals
the otocysts of Gonioncinns as tiny bubbles, each of which usually
encloses a single spherical object. This last mentioned body has
usually been considered as the otolith, though it probably com-
prises the entire sensory mechanism of the organ as indicated in
the introduction to this paper. Detailed structure of this body is
shown in Fig. 2, Plate I., which was drawn from a 15^ section.
This figure shows only the spheroid pendant from the wall of
the primary vesicle by the minute primary pedicel ( />/>)• The
heavily nucleated cells which comprise the wall of the secondary

MORPHOLOGY OF OTOCYSTS OF GONIONEMUS.

291

vesicle (sv) completely envelope the secondary pedicel (sp).
The true otolith (o) is shown within the secondary vesicle. In
the section from which this drawing was made a small tangential
slice such as is shown in Fig. 3, Plate I., had been cut from the
surface of the spheriod thus disclosing the cavity of the sec-
ondary vesicle. No sensory hairs or ridges were demonstrated.
Perkins ( 1902 : 786) also calls attention to the absence of such
structures, though his observations were probably confined1 to
the walls of the primary vesicle. In sections a thin tangential
slice from the surface of the spheroid. Text Figure 2, may at

TEXT FIGURE 2. Tangential slice only partly removed from wall of
spheroid.

first glance remind one of the projections figured by Hertwig
(1879: 183, Figs, i and 2) for Cannarina.

The primary pedicel measures 0.043 nim- m diameter and 0.069
mm. in length. At its proximal end a single large nucleus is
found. This pedicel is very delicate for in sectioning, it, together
with the spheroid, is often torn loose from the cyst wall. In
other cases only the spheroid is torn away from the distal end of
the pedicel and may be found elsewhere on the slide. Goto
(1903: 8) records encountering the same difficulty in the prepara-
tion of sections.

The secondary pedicel measures 0.115 mm. in length. Its
diameter near the middle is about 0.023 min. but the distal end is

L. J. THOMAS.

expanded to about 0.043 mni- Two nuclei of apparently fixed
relationships are found in the secondary pedicel, one near the
proximal and the other near the distal end.

The cavity of the secondary vesicle is normally spherical in
form though it is capable of some distortion. The diameter is
about 0.138 mm. The contents of this cavity take stains so
lightly and so evenly that there is strong evidence that only a
fluid is present. Two or three concentric rings are observable in
the spherical otolith contained within the secondary vesicle (Figs.
4-5, Plate I.). There is considerable variability in the size of
the otolith. Some of the largest measured 0.069 mm- m diameter.
The fluid in which the specimens were preserved was distinctly
acid, consequently any calcareous deposits that might have been
present in the otolith had been destroyed, leaving only the sup-
porting structures. Perkins (1902: 786) refers to the otolith as
" a calcium salt deposit in an organic matrix." In the specimens
examined the otolith had no fixed position within the secondary
vesicle but was apparently free to move about in the fluid filled
space. The entire sensory mechanism as ordinarily described for
an otocyst is thus contained within the confines of the spheroid.
In the light of this morphological evidence it might be easily pos-
sible that the destruction of the primary vesicle need not appre-
ciably impair the functioning of the organ.

Murbach (1903: 206) in experiments upon the function of the
otocysts of Gonioncmus collapsed the " otocysts " by thrusting
them with a fine needle. In all probability the injury inflicted
did not extend beyond the collapsing of the primary vesicle. Ac-
cording to his statements the specimens continued to act normally
after this treatment. Upon the results of these experiments and
upon the behavior of a single individual from which the otocysts
were excised he based his conclusion that the otocysts play no
important part in establishing the equilibrium in Gonionemus.
The more or less normal behavior of the much mutilated indi-
vidual from which the otocysts were cut is not readily explain-
able. On the other hand, in view of the fact that morphologically
the entire sensory mechanism of the otocyst seems to be confined
to the secondary vesicle there is little reason to expect serious

MORPHOLOGY OF OTOCYSTS OF GONIONEMUS. 293

interference with the functioning of that organ when only the
protective primary vesicle is collapsed.

My observations upon specimens of G. rcrtcns and of G. mur-
bachii have failed to disclose any essential difference in the de-
tails of structure or in the general orientation of the otocysts of
these two species.

RELATIONS OF OTOCYSTS TO BELL MARGIN.

Because of the reliance placed upon the works of previous
investigators and writers on the subject, Mayer's "Medusae of
the World " includes several erroneous statements and incorrect
figures of the otocysts of Gonionemus. Few authors, with the
exception of Mayer, have attempted to describe definitely the
location of the otocysts with reference to their position on the
margin of the bell. It seems probable that his description is
based chiefly on Perkins' publication (1903: 786) which has been
extensively quoted by Mayer though his observations upon the
otocysts are very misleading. In Plate 34, Fig. 19, Perkins has
reproduced a drawing by Professor Brooks which shows the
otocysts as external projections, from the margin of the bell
between the bases of the tentacles. Further, the drawing of
the "radial transverse section" of the bell (Fig. 25 of his same
plate) confirms the impression of their external location. His
explanation of their origin is as follows :

In the case of the sensory clubs, the endodermal tissue of the circular canal
grows down in a plug into the ectodermal tissue of the bell margin. This latter
becomes closely applied to the outside of the plug, as a thin investing epithe-
lium, and it also spreads out in a thin lamella over the inner surface of the
capsule which appears in the ectoderm of the developing club.

Iii his summary Perkins (1903: 789) says, "sense organs appear
at determinate points on the bell margin." The work in this
article on sense organs seems to be principally that of Professor
W. K. Brooks whose observations were apparently accepted by
Perkins without attempt at verification.

Mayer (1910: 341) in his synopsis of the genus seems to have
•'ncorporated the foregoing incorrect observations bodily for he
Defers to the otocysts as " lithocysts external." On page 342 of
the same work he again states that there are " numerous ex-

294

L. J. THOMAS.

ternal lithocysts upon the bell margin between the tentacles." In
characterizing the genus Cubaia (page 351), he states that there
are "lithocysts projecting outward as in Gonionemns". . .
" this genus is closely related to Vallentinia Brown 1902 but in
Vallcntinia the lithocysts are enclosed and on the inner side of
the margin, whereas in Cubaia they are external and on the lower
side of the margin between the tentacles." The plates of the
same monograph are just as confusing as the foregoing descrip-
tions in the manner in which the otocysts are located. Otocysts
are shown as distinct projections from the external margin of the
bell in Fig. i of Plate 45, though Fig. 2 of the same plate and
Fig. 3 of Plate 46 correctly portray them embedded in the meso-
glea as outgrowths from the ring canal.

The location of the otocyst within the mesoglea and the orienta-
tion with reference to other structures is shown in Fig. I,
Plate I.

RELATIONS OF OTOCYSTS TO TENTACLES.

Hargitt (1910: 249) referring to the location of the otocysts,
says : ' Normally they should occur in somewhat symmetrical
order between the bases of the tentacles. This, however, is rarely
the case." Mayer (1910: 342) gives the number of tentacles for
specimens of G. vert ens 15 mm. in diameter as 60 to /o but it
appears that his count is too low. In the following table are
given the data regarding the numbers of otocysts and tentacles in
part of the material studied in the present investigation.

TABLE I.

RELATIVE NUMBERS OF OTOCYSTS AND TENTACLES IN GONIONEMUS.

Specimen Number.

Diameter of Bell After
Clearing.

No. of Tentacles.

No. of Otocysts.

G. vertens
I

14.5 mm.

01

76

2

14

86

77

3 .

14

00

72

4- .

12

84

60

?

II

04

7?

6.

II

IOI

78

7

IO

04

80

8

Q

70

6l

Q

8?

C7

IO

8

70

V
60

G. murbachii
i

48

61

2 . .

10

6>

70

MORPHOLOGY OF OTOCYSTS OF GONIONEMUS. 2Q5

With reference to the tentacles these organs display no abso-
lutely fixed relationship. Though they usually alternate with the
tentacles they may occur in pairs between two adjacent tentacles
or two or more tentacles in continuous sequence may have no
intervening otocyst.

In G. vertens, nos. 6 and 7 had three paired otocysts between
adjacent tentacles ; no. 7 also had two abnormal otocysts with two

TEXT FIGURE 3. Abnormal otocyst with two sensory spheroids within the
same primary vesicle.

sensory spheroids in each capsule Text Figure 3. Hargitt (1901 :
249) has noted this last mentioned variation and has called
attention to its relative infrequence. Specimen no. I of G. mur~
bachii had thirteen paired otocysts and no. 2 of the same species
had eight.

SUMMARY.

1. There are two distinct vesicles in the otocyst of Gonloncnius.

2. It seems probable that the conspicuous primary vesicle has
chiefly a protective function, enclosing the essential sensory
mechanism.

3. The secondary vesicle contains the otolith within a fluid-
filled cavity.

4. No sensory hairs or ridges have been demonstrated.

5. Puncture of the primary vesicle as practiced by Murbach and
others probably does not impair the essential sensory mechanism.

6. The otocysts of G. zfcrtcns and of G. uiurbachii are in close
proximity to the periphery of the ring canal, their capsules im-
bedded in the mesoglea so that no portion of them protrudes
beyond the margin of the bell.

7. The number and arrangement of otocysts and tentacles in
Gonlonenius is variable.

296 L. J. THOMAS.

REFERENCES.

Agassiz, A.

'62 Contributions to the Natural History of the United States. Vol. 4,
Boston.

'65 North American Acalephae, 'Vol. 2, Cambridge.
Bigelow, H. B.

'09 The Medusae. Mem. Mus. Comp. Zool. Harvard College, Vol. 37.
Brooks, W. K.

'86 Life-History of the Hydromedusae. Mem. Boston Soc. Nat. Hist., Vol.
3: 359-430.

'95 The Sens-ory Clubs or Cordyli of Laodice. Jour. Morpnol., Vol. 10:

287-304.
Claus, C.

'78 Ueber Charybdea Marsupialis. Arb. Zool. Inst. Wien, Vol. i: 221-276.
Goto, S.

'03 The Craspedote Medusa Olindias and Some of its Natu.-al Allies. Mark

Anniversary Volume : 3—25.
Haeckel, E.

'81 Monographic der Medusen, Vol. 2. Jena.
Hargitt, C. W.

'oo Variation Among Hydromedusae. BIOL. BULL.. Vol. 2: 221-255.
Hertwig, 0. u. R.

'78 Das Nervensystem und die Sinnesorgane der Medusen. Leipzig.
Mayer, A. G.

'09 Medusae of the World, Vol. 2, The Hydromedusae. Carnegie Inst.

Washington, Publication No. 109.
Murbach, L.

'03 The -Static Function in Gonioneiuus. Amer. Jour. Physiol., Vol. 10:

201—209.
Murbach, L., and Shearer, C.

'03 On Medusa from the Coast of British Columbia and Alaska. Proc.

Zool. Soc. London, Vol. 2 : 164-189.
Nutting, C. C.

'99 The Hydroids of the Woods Hole Region. Bull. U. S. Fish Com., Vol.

19 : 325-386.
Perkins, H. F.

'03 The Development of Gonionemns inurbachii. Proc. Acad. Nat. Sci.
Philadelphia, Vol. 54 : 750—790.

298 L. J. THOMAS.

EXPLANATION OF PLATE.
SYMBOLS.

c, cavity of secondary vesicle, p p, primary pedicel,

e c, ectoderm, r, ring canal,

e ii, endoderm, s, spheroid,

e x, exumbrella, s I, supporting layer,

m, mesoglea, ^ p, secondary pedicel,

n, nerve ring, s 11 , subumbrella,

n c, nettling cells, j v, secondary vesicle,

o, otolith, 7', velum.

PLATE I.

The otocyst of Gonionemus rertens. All drawings were made with the
camera lucida. The projected scale indicating magnification in each instance
has the value of o.oi mm.

FIG. i. Section through bell margin showing general location of the
otocyst within the mesoglea and its orientation with reference to ring canal.

FIG. 2. Longitudinal section through the spheroid of an otocyst showing
histological details.

FIG. 3. Tangential slice from the wall of the spheroid similar to the sec-
tion removed from the front surface of the wall of the secondary vesicle in
the foregoing figure.

FIG. 4. Longitudinal section through an otocyst to show the concentric
rings within the otolith.

FIG. 5. A spheroid with a portion of the wall of the secondary vesicle
removed.

BIOLOGICAL BULLETIN VOL. XL.

PLATE I.

L. J. THOMAS.

A SYMBIOTIC FUNGUS OCCURRING I> THE FAT-
BODY OF PULVINARIA INNUMERABILIS RATH.1

CHARLES T. BRUES AND RUDOLF W. GLASER.

During the late winter and spring of 1920 the present writers
became interested in the supposedly symbiotic organisms which
occur in various scale insects, hoping that they might be able to
propagate some of them in artificial cultures and learn something
of their physiological activities. From knowledge gained thus,
it seemed probable that they might be able better to determine
whether such organisms exist in the insects as true symbionts,
as mere commensals or as innocuous parasites.

As is well known through the investigations of several workers,
the entrance of the symbionts2 into the egg of scale insects can be
readily followed as well as their behavior during embryological
development. A good account of this has been given by Shinji
('19) who also includes a summary of the previous work of other
authors.

In the nymphal or full-grown scales of many species it is more
difficult to find and interpret the symbionts and it seems probable
that in some cases they must either become very much reduced in
numbers, very highly modified, or perhaps reduced to minute and
unrecognizable spores or granules.

Several workers who have recently examined the symbionts of
Coccids (e.g., Buchner '12, Sulk '10, Teodoro '18) regard them
as yeasts (Saccharomycetes), although Berlese ('06) referred one
species to the genus Oospora, one of the Hyphomycetes.

Before beginning our work with Pulvinaria innumerabilis, the
cottony maple scale, we examined several other species of Coccids,
but were unable successfully to cultivate the symbionts from
these, with the possible exception of one of the pine scales,

1 Contribution from the Entomological Laboratory of tht Bussey Institu-
tion, Harvard University, No. 176.

2 We have used this name as a convenient designation already in current
use and will discuss its appropriateness on a later page.

299

3OO CHARLES T. BRUES AND RUDOLPH W. GLASER.

Cliionaspis pinifolicc Fitch. From the latter we isolated an or-
ganism, perhaps related fo Oospora, but we could not secure it
with sufficient regularity to satisfy ourselves that it was really the
symbiont. and not a contaminating species of microorganism of
which one always encounters numerous species in work of this
kind. Just how closely related the symbionts of various Coccids
may be, must remain a matter of doubt until a considerable num-
ber have been carefully investigated, and preferably cultivated
also, but our own observations lead us to think that more than
one type of organism will be found after careful, systematic
study.

In the following brief review of literature we have considered
only such contributions as appear to bear directly on symbionts
quite probably closely related to the form with which we have
worked.

The first reference to the occurrence of symbiotic organisms in
Coccidse is that of Leydig ('54) who found discrete, lanceolate
bodies which he believed to lie free in the lymph of Coccus (now
Lecanium) hesperidum. He described them as 4^ in length,
multiplying by buds which do not separate till they have attained
the size of the parent cell. Neither at this time, nor in 1860 in
his contribution to the development of the Daphnids, did he
realize their significance. From the rather large size and method
of multiplication, these are probably similar to the organisms
found in Puh'inana, which also appear in freshly mounted smears
as though they were free in body fluid. Putnam ('79) studied in
considerable detail the biology, anatomy and development of
Pulvinaria innumcralnlis in Iowa. In connection with a section
devoted to the contents of the ovaries he gave an account of the
organisms with which we deal in the present paper. His observa-
tions were so carefully made that we have thought it worth while
to include the following resume. On opening a female at any
time from October to May, five classes of bodies are set free, all
of them apparently associated with the development of the eggs.
These are : First, a clear protoplasmic liquid ; second, clear
spherical globules 10 ^-30 /x in diameter, lighter than water, r.nd
not taking the ordinary aniline stains. He was undoubtedly cor-

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 30 1

rect in believing these to be yolk and fat globules. Third,' ex-
ceedingly minute, apparently spherical bodies, heavier than water
and staining with cosine. Putnam thought that these might be
bacteria although he suggests that they may be comparable to the
blood-disks [erythrocytes] of vertebrates or that they may be
stages in the development of the fourth class of bodies which he
next considers. That all of these suppositions are probably in-
correct will appear from our account on a later page. Fourth,
small oval bodies 3/^-5^ in diameter and io/x long and heavier
than water. These represent the organism which we have studied
but Putnam was unable to decide whether they were spermato-
phores or whether they corresponded to the pseudonavicella:
observed by Ley dig in Lccanium, as had been suggested to him
by Dr. E. L. Mark.1 A fifth class of bodies observed were the
small incompletely formed eggs.

There can be no doubt that Putnam found the fungus with
which we have worked as his description and figures make this
point very clear. As he found it in all cases, it is further clear
that the symbiont enjoys a wide range since his observations were
made on specimens collected in the middle west and our own on
material from eastern Massachusetts.

Metschnikoff ('84) found in a crustacean, Daphnla magna, a
fungus which he called Monospora bicuspidata. This he re-
garded as a parasite, but recent developments in the study of
apparently similar organisms in insects, suggests that these Crus-
taceans should be reexamined.

Moniez ('87) described a fungus which he called Lccaniascus
polymorphus, occurring in the scale insect, Lecan'mm hesperidum.
He refers to Leydig's 1854 paper, mentioning the fact that
Lecaniascus is evidently the same as Leydig's pseudonavicellae.
Moniez speaks of the organism as a parasite, and he found it in
all specimens, both young and old, of the Lccanium that he ex-
amined. He described the isolated cells as 4-5 ^ in length, and
found mycelia attaining a length of 50-60 p.. Some doubt is cast
upon this author's conclusions by Vejdovsky ('07) who suggests
that Moniez may have seen two microorganisms, one represented

i Mark ('77) does not consider these bodies, however, in his paper on the
anatomy and histology of the Coccidse.

3O2 CHARLES T. BRUES AND RUDOLPH W. GLASER.

by the single cells and another by the mycelia and asci, the latter
perhaps Alternaria tennis, a parasitic fungus that attacks various
Coccids of the genus Lccanhnn.

Lindner (1895) found in an European scale insect (Aspidiotits
ncrii) a yeast-like organism which he regarded as related to
Saccliaroin\ccs aplculatus. By crushing one of the insects be-
tween a slide and cover-glass he observed large numbers of the
yeasts, both between the small masses of fat-body and actually
in the adipocytes. The organisms he described and figured as
usually very long, pointed at one end, or lanceolate, sometimes
joined in pairs by their acute tips, and frequently budding after
the manner of yeast cells. At that time Lindner was unable to
cultivate the organism although he evidently regarded it as a.
parasite, naming it Saccharomyces aplculatus, var. parasitus. He
found it forming a mass at the posterior pole of the egg and made
an ingenious explanation for its presence there, suggesting that
one of the pointed tips perforated the egg and gave off a bud
which then multiplied to form the mass or mycetome.

A later paper by Lindner ( '07) which appeared in the Woch-
cnschrift filr Braucrcl we have not seen, but from published
reviews of this, it appears that it contains nothing bearing on the
physiology or systematic position of his yeast-like organism.

Berlese ('06) studied in detail an organism which he found in
Ceroplastcs ruscl, and was successful in growing it on artificial
media. From Berlese's account, it appears that the fungus,
which he calls Oospora saccardlana is very similar to the one
described by us in the present paper. The yeast-like cells in the
Coccid vary in length from 4-12 //,, sometimes attaining a length
of 16-18/4 in early summer, agreeing in size and also in form
with those we have observed in Pulvinaria. In culture there is z
great similarity in the general morphology and development of
mycelia, and although Berlese gives few details concerning cul-
tural characteristics, at least one statement shows a striking dif-
ference between the two. He found that although the symbiont
from Ceroplastcs grows rapidly on gelatine media, that these arc
not liquified, while as will appear from our account, the Puh'l-
naria fungus exhibits a powerful liquefactive action on gelatine.

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 303

Concerning the distribution of the symbionts in the body of
Ccroplastes, Berlese gives no account, except to state that they
generally occupy the visceral cavity completely in all individuals,
in numbers estimated at from 60,000-70,000 cells.

Two other genera of soft scales, Kcnncs and Physokermes,
have been the subject of investigation by Sulc ('07) and Vejdovsky
('07). The former found two distinct symbiotic organisms in
two species of these Coccids, readily distinguishable from one
another on the basis of size and form. These he described as
representatives of a new genus, Kerminicold. Vejdovsky re-
garded them as Saccharomycetes in which opinion Sulc con-
curred. The microscopical structure of the symbionts was care-
fully described and is similar to that of the species in Pulvinaria
studied by us, although Kcnninicola evidently has a much more
prominent and discrete nucleus and fewer multinucleate cells ;
also the form of the cells is generally much more elongate.
Vejdovsky found the symbionts in the fat cells of the host in
large numbers and observed them freed in the haemolymph as a
result of a disintegration of the adipocytes which he believed due
to the activity of the included organisms. He, therefore, re-
garded them as parasites, but pointed out that their activities do
not affect the gonads of the host, nor the development of the eggs
in its body. They do, however, serve to break down the fat and
to consume the remaining tissues of the host's body, after which it
remains a shell for the protection of the now fully developed
nymphal Kcrmes.

In a later paper, Sulc ('10) gives a more extended account of
the similar organisms of a number of Homoptera and speculates
at considerable length upon the relations between the insects and
fungi. After more extended study, his ideas have been consider-
ably modified and he has come to regard the microorganisms as
essentially symbiotic in their association with their host insects.
He suggests that the production of enzymes by yeasts (for he
still refers the symbionts to the Saccharomycetes) must be con-
sidered in any interpretation of their physiological relations to the
insects. Apparently he made no attempts at cultivation in vitro.

Pierantoni has considered the symbionts of certain Coccidze in

304 CHARLES T. BRUES AND RUDOLPH W. GLASER.

several papers of which that of 1910 in the Zoologischcr Ansc:gcr
is of greatest interest in the present connection. In Iccr\a pur-
chasl, he traced the entrance of the organisms into the egg of the
scale insect and its subsequent behavior through the formation of
the polar mass in the egg to the development of the mycetocyte
in the larva. He found that the individual symbionts of this
species were at first round or oval, and not noticeably elongated,
and that later during embryonic development and at the time of
hatching they became quite inconstant in form, varying from
rounded or oval to much elongated and frequently strongly
curved cells, all, however, of about the same diameter. These
long forms may fragment, each piece becoming a new individual,
while the short ones commonly divide by fission into equal parts.
Occasionally, however, Pierantoni found cells multiplying by
budding in the body cavity of the host, and more rarely in the
mycetome.

On a gelatine medium, with high sugar content (20 per cent,
saccharose) he was able to cultivate a yeast-like organism from
the mycetocytes. These developed after four days' incubation as
colonies that are described as small spheres in the gelatine which
develop a sort of finger-like process which becomes prolonged
toward the surface of the gelatine and then emerges projecting
above it in the form of a finger, or with the base enlarged and
pyriform. The individual organisms were of yeast-like form with
buds more or less developed. It thus appears that the organism
obtained in culture by Pierantoni differed greatly in form, size,
and method of multiplication from the organisms in the insects,
as will be shown later, the cultures obtained by us from Puh'i-
naria innumerabilis exhibit no such remarkable distinctions, in
morphology and reproduction, from the cells in the host insect.
Although he makes no mention of the development of mycelia in
his cultures, the form of the colonies indicates without question
that such must have been present, and that if the organisms in
the mycetome were actually those obtained in culture, the
symbiont of Icer\a is a true fungus.

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 305

THE SYMBIONTS OF THE HALF-GROWN PULVINARIA.

In early April the overwintered cottony maple scale-insects are
partly grown and may be found attached to the bark of small
twigs on the food plants, which consist of various maples and a
few other wroody plants. At this time they are nearing the end of
their period of hibernation and do not yet exhibit any active
growth.

If a specimen in this condition be crushed on a slide in normal
salt solution, the symbionts may be readily seen free in the liquid.
They are heavier than the medium and fall next to the slide, thus
separating from the released fat globules which accumulate above,
against the cover-slip. In such a preparation all the organisms
appear to be in the liquid, as those in the fat cells are not readily
discernible on account of their hyaline nature. This, no doubt,
accounts for the statements that the symbionts occur in the lymph
rather than in the tissue.

In sections it is evident, however, that the organisms are absent,
or at least very nearly so, from the blood and that they are very
generally distributed through the fat body, imbedded in the adi-
pose cells. They are usually spaced in a quite regular way show-
ing that they migrate or at least change their position in the cells
subsequent to multiplication. The density of distribution is well
indicated in the drawing (Fig. I, A) which is made from a section
of 6 /x. in thickness where all of the symbionts present have been
sketched. The photograph on Plate I., Fig. i, is made from a
typical cross-section through an entire insect, in which the sym-
bionts appear as minute oval dots. CM Plate I., Figs. 2, 3 and 4,
are reproduced several small areas of the fat-body viewed at
higher magnification with their symbiont inclusions. Sketches
of a few still more highly magnified symbionts are shown in Fig.
i, B. Here it will be seen that they are extremely variable in size
and shape, but always quite distinctly oval in form with one pole
more acute and the opposite one more rounded. They vary in
length from 1 0-16.7 /x by 5-6.5^ in width, with an average size
of 10-12.5 /A by 5.7^. Budding forms are frequently present, the
bud developing at the narrow pole and separating either as a small
oval, or more rarely rounded, cell. The buds at the time of sepa-

306

CHARLES T. BRUES AND RUDOLPH W. GLASER.

ration are of the same general shape as the cells from which they
originate, but much smaller, varying in length from 6.2-6.7 /x.
Some buds are nearly round, in which case they separate when
considerably smaller, about 3.7 /x in diameter.

.••••,-•• , - SE&F&aB

ill ' ' ^^^SKM

• -- ' -' * '" ^" • '- ' ~^<- ' • "' t^*- • "•"•" '• • ' -'-(•', L\'---L*l ' ^ —k/7'

T; ,,

FIG. i. a, Section of fat body showing symbionts in situ in overwintered
nymphal Pulvinaria, X 260 ; b, isolated symbionts, stained with methylene blue
and eosin, X 1,000.

The internal structure of the symbionts shows very little uni-
formity. Sections of tissue fixed in Zenker's solution and stained
with methylene blue and eosin show one or several rather distinct
deeply stained portions that resemble nuclei. When one is pres-
ent it is usually central, when two or more are present, they are
separated rather evenly from one another and from the cell wall.
The remaining protoplasm shows irregular denser streaks and
spots of irregular size, with usually one large, several smaller, or
one large and one small, vacuole, with generally a number of
minute clear spots that can be seen only after very close examina-
tion. We have tried to bring out other details of structure by dif-
ferent methods of fixation and stajning but without much success.
Smears fixed by such methods as drying and submersion in abso-
lute alcohol and subsequent staining with Giemsa's stain, Hanson's
Protozoan stain and carbol-fuchsin, give quite similar pictures to
those in sections which are disappointing from a cytological stand-
point. Evidently the symbionts are of very delicate consistency.
No chains of symbionts occur at this time (early April) except for
the single attached buds, and no masses of contiguous ones are to
be found.

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 307

CHANGES IN THE SYMBIONTS AFTER HIBERNATION.

Our observations on these are very fragmentary as our available
time at this season was occupied with the cultural experiments
described below. It appears quite certain, however, that the oval
symbionts of the early spring Pnhnnaria nymphs undergo further
multiplication and morphological changes in late April, and dur-
ing May. A cursory examination of specimens about May 10
showed the symbionts in groups sometimes forming short strings,
somewhat similar to the typical mycelium which was obtained in
cultures. The picture at this time indicates cmite clearly the
fungous nature of the yeast-like organisms just described, found
before spring growth begins.

%

ISOLATION OF THE FUNGUS FROM PULVINARIA.

Our first attempts to cultivate in artificial media the yeast-like
cells present in Puh'inaria were made in early April. The over-
wintered, partly grown scales were brought into the laboratory
still attached to the maple twigs on which they occur. The scales
can readily be detached from the twig by means of a sterile
needle and allowed to fall upon a sterile microscope slide. The
scales thus removed, were treated in several ways as experience
had taught us that material of this sort is very apt to be contami-
nated on the surface by various microorganisms. Some were
treated with 85 per cent, alcohol for a few minutes, others rapidly
passed through the flame of a Bunsen burner, and others were
used without treatment, except to avoid contact with any un-
sterile object. Our first media were potato agar and " sugar
agar " which consisted of potato agar to which varying amounts
of maple sugar (2l/2 per cent.- 20 per cent.) had been added.
These tubes are readily inoculated by crushing the scale insect
between two microscope slides and streaking the agar with a loop
dipped in the body-juices thus extracted. It is also easy to drop
the whole insects into culture tubes and then crush them on the
agar by means of a dissecting needle. From a large series of
tubes inoculated according to these methods, nearly one half
showed after three days a good growth, white in color, spread-
ing on the surface of the media. These colonies appeared to de-

3o8

CHARLES T. BRUES AND RUDOLPH W. GLASER.

velop almost equally well in the greatly diverse concentrations
of sugar and in the plain potato agar. A microscopic examina-
tion at this time showed that nearly all the tubes in which growths
occurred contained the same microorganism, and that only a few
were contaminated by molds (PemcilUum) and bacteria. The
abundant species showed large numbers of budding yeast cells
like those in the living scale insects and the development of my-
celium as well showing that the symbiont was a fungus and not a
yeast as one might otherwise be led to believe from a study of the
living insects, at this season of the year when only single budding
cells occur in the fat body.

Several of the colonies thus obtained were plated, found to be
pure and sub-cultures were then made from these which have
furnished the material for the description and cultural characters
detailed below. Although the morphology of the organism in the
living insects and in the cultures and the fact that it was recovered
in such a large proportion of the cultures, left little doubt as to
the identity of the two, we undertook some serological tests to
corroborate if possible the conclusion based on morphological
data.

For this purpose, two rabbits were secured, inoculated with
bouillon cultures of the fungus, and serum from each was tested
with the cultures and also with the organisms in the living scale
insects. The following table shows the doses used and the reac-
tions of the rabbits.

Date.

Amount of Culture.

Weight.

Rabbit A.

Rabbit B.

April 13

3 c.c.
5 C.C.
5 c.c.
6 c.c.
6 c.c.

2185 gms.
2180 gms.
2080 gms.
2110 gms.
2070 gms.

1905 gms.
1910 gms.
IQIO gms.
1895 gms.
IQI=; gms.

April 1 6

April 20

April 24

Anril 20 .

One week later some blood was withdrawn from each rabbit
and serum prepared. A precipitin test with the culture gave a
positive reaction after four hours, and agglutination was very
pronounced after one and one half hours as examined under the
microscope. On account of the impossibility of securing suf-
ficient material from the living scales, it was possible to try with

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS.

them only the agglutination test. With these the reaction was
not so pronounced as with the yeast-like forms in culture, but
nevertheless distinctly positive. Unfortunately by the time the
animals had been immunized (early May) the number of yeast-
like cells in the insects had decreased and the reaction could not
be so readily observed as in the cultures, or so well as it might
have been several weeks earlier in the spring when the insects
contained innumerable, separated, oval cells. The data from the
rabbit experiments has, however, convinced us that there can be
practically no question that the organism cultivated is actually
the one present in the insects. Furthermore, since we have never
failed to observe it in living scales from this locality, and since
Putnam (v. antca, p. 301) found it invariably present in Iowa, it
is undoubtedly present regularly in Pnlvinaria inmimerabilis.

CULTURAL CHARACTERISTICS OF THE FUNGUS.

As stated at the outset, we wished to grow the symbiont on arti-
ficial media, not only to describe it adequately, but to determine
as completely as possible its physiological activities. In order to
do this, sub-cultures from the original isolations were planted
upon various media, such as are in general use by bacteriologists
and mycologists. From these, the following observations were
made.

Growth on Solid Media.

Potato Gelatine Colonies. — Growth rapid ; after 72 hours, cot-
tony with flocculent elevated center and filamentous edge, diam-
eter of center I mm., width of margin I mm. Liquefaction cup-
shaped.

Nutrient Gelatine Colonies. — Growth slow ; after 72 hours,
rounded, with central elevation. Diameter 0.3 mm., with roundly
lacerated edge. Liquefaction cup-shaped.

Potato Agar Colonies. — Growth rapid; after 72 hours, fila-
mentous, ciliate (sub-surface) or rounded (surface). Disk when
present, smooth ; elevation convex ; edge of round colonies smooth,
that of irregular colonies radiately filamentous or ciliate. Internal
structure finely granular. Diameter 1.5-2.5 mm.

Nutrient Agar Colonies. — Growth slow ; after 72 hours, round,

3IO CHARLES T. BRUES AND RUDOLPH W. GLASER.

with smooth surface and convex elevation. Edge smooth ; inter-
nal structure finely granular, with irregular central core. Diam-
eter 0.6 mm.

Potato Gelatine Stab. — Liquefaction begins in 48 hours ; growth
best at tip, the line of puncture filiform ; liquefaction at first napi-
form, becoming stratiform, surface umbonate. Medium un-
changed.

Nutrient Gelatine Stab. — Same, but liquefaction proceeds more
slowly and is more nearly crateriform.

Potato Agar Slant. — After 72 hours growth is abundant, spread-
ing, densely rhizoid, convex. Color white, opaque, surface glis-
tening, smooth. No odor, consistency viscid. Medium un-
changed.

Nutrient Agar Slant. — After 72 hours growth is moderate,
beaded, with a few rhizoid colonies where medium is thin. Sur-
face smooth, glistening. Color white, opaque. No odor, con-
sistency butyrous. Medium unchanged.

Locke's Agar Slant. — Growth like other agar slants, scanty,
beaded, many rhizoid colonies, consistency butyrous. After a
much longer incubation (four weeks) there is not a very heavy
growth, but it is still more or less beaded and highly rhizoid on
the sides and extending deep into the agar.

Molisch's Agar Slant. — After 72 hours growth abundant,
much like that on potato agar. Consistency very ropy. No pig-
ment, even in old cultures.

Potato Agar (with Yeast1) Slant. Growth after 72 hours
abundant, like that on potato agar, but with fewer rhizoids.

Potato (Pieces). — Colonies raised, slightly yellow, growth
good. There is a trace of ammonia.

Growth in Liquid Media.

Nutrient Bouillon. — After 48 hours growth good; after 72
hours, no surface growth, no clouding and no odor, but with an
abundant viscid sediment. No hydrogen sulphide is produced.

Locke's Solution. — After 48 hours not very much growth;
after 72 hours, no surface growth, no clouding and no odor, but
with an abundant viscid sediment. After a much longer incuba-

1 Made by adding 2 per cent, of thoroughly crushed bread yeast.

FUNGUS OCCURRING IN PULVINARIA IN NUMERABILIS. 3! I

tion at room temperature (four weeks), the growth becomes very
flocculent and adheres to the surface of the glass where it is
finely dotted with very dark, pigmented spots.

Moli-sctis Solution. — After 48 hours rather good growth ; after
72 hours, no clouding or surface growth, but with very abundant,
slightly viscid sediment, no odor.

Sugars. — We have grown the organism in six sugars : lactose,
dextrose, mannite, saccharose, levulose and maltose. None of
these, however, furnish any differentiating characters ; in all
there is good growth without the formation of gas, but with
heavy viscid sedimentation, and after prolonged incubation, the
development of a distinct surface film.

Milk. — Growth is abundant, and after four to six days incuba-
tion the milk begins to clear at the top, sediment collecting in the
lower half of the liquid. Litmus milk becomes distinctly red at
the time of clearing. After about a week, the liquid becomes
whey, with a sediment at the bottom of the tube.

Anaerobic Media. — We have not been able to cultivate the or-
ganism under anaerobic conditions.

PRODUCTION OF ENZYMES.

Protease. — Gelatine is rapidly liquefied. Milk cultures, tested
after 20 days, give a positive reaction with MgSO4, NaOH and
CuSO4, showing the presence of peptones.

Lipase. — After both 20 and 30 days' incubation, cultures in
either whole or skimmed milk give positive reactions. There is
a strong odor of butyric acid. Ethyl butyrate is also decomposed
with the formation of butyric acid. The 3O-day culture was
tested with pyrogallol and stannic chloride and gave a positive
reaction also.

Diastase. — Bouillon cultures after ten days' incubation were
treated with starch paste at incubator temperature for 48 hours ;
after this, Fehling's solution was reduced, demonstrating the
presence of sugar.

MORPHOLOGY OF THE CULTIVATED FUNGUS.

The cultures show during the first few days only yeast-like
budding cells like those seen in the early spring in the fat-body

312

CHARLES T. BRUES AND RUDOLPH W. GLASER.

of the insects. These vary considerably in size, ranging from
6-i6;u in length and from 3-9^ in width. They are thus more
variable in size in culture than in the insect and generally more
elongate. The maximum length is about the same, but there are
more smaller cells in culture, due no doubt to the fact that during
rapid development the buds separate when less fully developed
than in the insect. The internal structure when stained, appears
to be the same as that of the forms in the fat-body described
above.

After prolonged incubation on solid media the formation of a
distinct mycelium always occurs. This is at first wThite, but after
several wreeks, blackened spots sometimes become visible, due to
the development of pigment in the walls of certain groups of
cells. This occurs especially on potato-agar. In at least one
liquid medium, Locke's solution, the same blackened cells develop.

FIG. 2. Portion of mycelial growth of symbionts after prolonged incuba-
tion (10 days) in liquid bouillon medium. Magnified about 400 diameters.

The mycelium (Fig. 2) is branched and of quite irregular
form. The larger hyphse measure from 6-i$p. in diameter,
broad and narrow cells frequently alternating or with one size
interpolated in series of the other. Some cells, usually single
ones or pairs, more rarely several in succession are heavily pig-

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 313

merited and appear very dark in fresh material ; a pair of these
are frequently rather closely fused to form an oval body. The
contents of these dark cells are highly granular, with the proto-
plasmic mass clearly separated from the cell wall by a hyaline
layer. Toward the tips the hyphse are usually much more slender,
pale in color and with only scattered granules in the protoplasm.
The method of branching is entirely dichotomous, many lateral
branches along the larger hyphse consisting of but a single cell
although both near the tips and on the larger hyphae there are
many long branches which again subdivide. Very rarely there is
an anastomosis of the finer apical branches formed by lateral
prolongations of the cells. In preparations we have had difficulty
in recognizing the conidiophores, but free conidia are present in
old cultures. They measure from 8-io/x in length and are
broadly oval in form, very nearly transparent and without color.

SYSTEMATIC POSITION OF THE CULTIVATED FUNGUS.

We have been unable to deal with this matter in a satisfactory
way owing to our un familiarity with cryptogamic botany and
must leave it for consideration by an experienced mycologist. It
is very evident that the symbiotic organism in Pulvinaria cannot
be regarded as a Saccharomycete, although its morphology and
method of multiplication in the insect does not preclude such an
assumption. Indeed, the symbionts that have been observed in
other Coccids and in most other Homoptera as well have usually
been regarded as yeast-like organisms and commonly referred
to the genus Saccharomyces or to new genera located in the same
group of plants. Berlese ('06) has placed the organism which
he cultivated from Ccroplastcs in the genus Oospora, thus recog-
nizing it as a true fungus, but hitherto, with the possible exception
of Pierantoni ('10) no one else seems to have been successful in
growing in vitro any of the symbionts of coccids.

The species which we have obtained from Pnh'inaria seems to
be quite similar to the one described and figured by Berlese from
Ceroplastcs so far as the general morphology of the yeast-like
cells in the coccid and the mycelial structure in culture. Neither
species, however, has been sufficiently studied to make a more posi-

3 H CHARLES T. BRUES AND RUDOLPH W. GLASER.

tive statement. Dr. O. F. Burger kindly examined some of our
cultures and has expressed the opinion that they probably repre-
sent a species of Dematium or a related genus. Such morpho-
logical characters as we have been able to make out agree well
with descriptions of this genus to which it may be tentatively re-
ferred.

PHYSIOLOGICAL ROLE OF THE SYMBIONT.

As stated at the outset, we have attempted to determine the
physiological behavior of the Pulvinaria symbiont in culture to
ascertain in what way it may affect the metabolism of the coccid.

Contrary to what occurs in the case of most yeasts, this or-
ganism produced no gas in media made from any of the sugars in
which it was grown. This is quite what might be expected as
the coccid tissues are undoubtedly rich in sugars and any organ-
ism producing gas in the presence of such substances could not
be tolerated in the body of the coccid.

On the other hand a diastatic ferment is produced in quite ap-
preciable quantities. Whether this bears any relation to the
metabolism of the coccid is not entirely clear. In the adipose
tissue and body liquids, starch is probably not present to any con-
siderable extent, although in the large quantities of plant sap in-
gested by the coccids there must be substances upon which this
ferment might act. It has been shown also by Biisgen ('91) that
certain modifications are produced in the tissues of the food
plants of Coccids at the point where the mouth setfe are thrust
into the plant. These modifications appear to be induced by se-
cretions actually injected into the plant tissue by the insects and
they may act in liquefying or in partly digesting already liquid
or semi-liquid material, before it is withdrawn by the insect. It
is quite possible therefore that a diastatic ferment might act in
two possible ways in aiding the digestion of the coccid. If freed
in the blood, it might either pass into the alimentary tract, there
to act upon ingested food, or it might be taken up by the salivary
glands to be later injected into the plant and thus act as an extra-
intestinal digestive agent. Such extra-intestinal digestion is
known to occur in several diverse insects, although in these cases
the ferments are no doubt elaborated directly by the salivary
glands.

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 3!$

We have also clearly shown that a proteolytic enzyme and a
lipase are produced abundantly by the Pulvinaria symbiont. The
possibilities for these to influence the digestion and metabolism
of the coccid are more diverse than those presented by the pres-
ence of diastase. Both, particularly the lipase, must act upon the
adipose cells in which the symbionts occur. We might therefore
suppose that they assist in the rapid breaking down of this tissue
at the time of maturity when the eggs of the Pulvinaria are rapidly
developed. That they may assist in digestion, either in the body
or through the agency of secretions injected into the plant is also
quite possible, although such indirect action must undoubtedly
be secondary if it occurs at all.

THE GENERAL NATURE OF THE RELATION BETWEEN SYMBIONT

AND COCCID.

The symbionts have gradualy come to be regarded rather gen-
erally as truly symbiotic organisms, although those who first
studied them naturally assumed that their presence indicated
some sort of parasitism. There are several reasons why it is dif-
ficult to believe that they are actually parasitic. In the first place,
not only in Puhnnaria, but in the other species in which they have
been found, they are universally present in all the individuals of a
species in approximately equal numbers. Many true parasites,
e.g., certain Nematode worms, the Protozoan parasites of human
malaria, etc., commonly appear with great frequency in the bodies
of their hosts, but their occurrence never includes all the individ-
uals of a host species, except at certain times and places where
parasitism is unusually heavy and assumes the form of an epi-
demic. In 'such cases also, the affected population is not in a
healthy condition and species so generally affected cannot be ex-
pected to represent ones well fitted to survive and become abun-
dant. Pulvinaria and other Coccids certainly cannot be placed in
such a category. On the other hand the presence of detrimental
parasites results in tissue changes or disturbances of metabolism
that can be recognized. Such can be seen in the behavior of the
fat-body in Pulvinaria and other genera ( Sulc, '11), but as has
just been said this is most readily regarded as beneficial rather

316 CHARLES T. BRUES AND RUDOLPH W. GLASER.

than detrimental as the fat is not broken down until the time that
it would normally disintegrate to supply nourishment for the de-
veloping eggs of the coccids.

Without any definite indication of pathological changes, it
seems impossible, therefore, to regard the universally present
symbionts as harmful parasites.

It has also been suggested that they may represent innocuous
or indifferent parasites and it is not so easy to distinguish between
these and true symbiotic or benign organisms from their effect on
the coccids. As a matter of fact it seems necessary to regard all
three as steps in an evolutionary process, harmful parasites in
their first association, later as innocuous ones and finally as true
symbionts. These will follow one another as the host adapts
itself to withstand or nullify any ill effects of the parasite until it
finally is able to utilize the products of the intruder to further its
own metabolic processes.

Thus it seems reasonable to regard these three types of asso-
ciation as not clearly distinct from one another, but as connected
by intergrades.

Since, however, there is good reason to believe that the pro-
duction of diastase, protease and lipase by the symbionts may
serve to benefit the coccids, the possibility of real symbiosis can-
not be excluded.

There is one point, however, which needs further study. By a
minute study of the changes in the tissue of the food-plant ad-
jacent to the proboscis of the feeding Coccid, it should be pos-
sible to gain much additional evidences upon the changes which
undoubtedly occur in such tissue. This we have not had oppor-
tunity to undertake. Why the disintegration of the fat-body is
delayed till the proper time in the life-cycle of the coccid also is
not clear. Since, however, changes in the vegetative character
of the symbiotic fungus are initiated in the late spring, it seems
probable that they may determine to some extent the quantities
of enzymes produced. On the other hand it is evident that the
coccid is able to inhibit any excess development of the symbiont
as the number of symbiont cells remains very uniform and never
seems to increase beyond certain bounds, quite a different condi-

FUNGUS OCCURRING IN PULVINARIA INNUMERABILIS. 317

tion from that obtained among pathogenic parasitic microor-
ganisms.1

This indicates a nice physiological balance between the coccid
and symbiont and is another reason for considering this a case
of true mutualism.

BIBLIOGRAPHY.
Berlese, Am.

'06 Sopra una nuova specie mucidinea parassita del Ceroplastes nisei.

Redia, Vol. 3, PP- 8-15.
Breest, F.

'14 Zur Kenntnis der Symbionteniibertragung bei viviparen Cocciden und

bei Psylliden. Arch. f. Protistenk., Vol. 24, pp. 263—276, 2 Taf.
Buchner, P.

'n Ueber intrazellulare Symbionten bei zuckersaugenden Insekten und
ihre Vererbung. SB. Ges. Morph. Physiol. Miinchen, Vol. 27, pp. 89—96.
'12 Studien an intracellularen Symbionten. I. Die intracellularen Sym-
bionten der Hemipteren. Arch. f. Protistenk., Vol. 26, pp. i— 116.
Taf. 12.
Biisgen, M.

'91 Der Honigtau, biologische Studien an Pflanzen und Pflanzenlausen.

Jenaische Zeitschr. Naturwiss., Vol. 25, pp. 339—428, Taf. 2.
Howard, L. 0.

'oo The Two Most Abundant Pulvinarias on Maple. Bull. Div. Entom.

U. S. Dept. Agric., n. s., No. 22, pp. 7-16.
Leydig, F.

'54 Zur Anatomic von Coccus hesperidum'. Zeits. wiss. Zcb'l., Vol. 5, pp.

i — 12, Figs. 6
Lindner, P.

'95 Ueber eine in Aspidiotus nerii parasitisch lebende Apiculatus-hefe.

Centralbl. f. Bakt., Vol. i, pp. 782—787, Figs. g.
'07 Das Vorkommen der parasitischen Apiculatus-hefe. Wcchenschr. f.

Brauerei.
Mark, E. L.

'77 Beitrage zur Anatomic und Histologie der Pflanzenlause, insbesondere

der Cocciden. Arch. Mikr. Anat., Vol. 13, pp. 31-86, Taf. 3.
Mercier.

'07 Recherches sur les bacteroides des blattides. Arch. f. Protistenk., Vol.

9, pp. 346-358, Taf. 2.
Metchnikoff, I.

'84 Ueber eine Sprosspilzkrankheit der Daphniden. Arch. f. path. Anat. u.

Physiol. u. f. klin. Med., Vol. 96, p. 177.
Moniez, R.

'87 Sur uti champignon parasite du Lecanium hesperidiim (Lcciniascns
polymorphns nobis). Bull. Soc. Zool. France, Vol. 12, pp. 150—152.

1 In this connection it is interesting to note that we found antibodies devel-
oped abundantly in rabbits immunized against our cultures of the Pulvinaria
symbionts.

3l8 CHARLES T. BRUES AND RUDOLPH W. GLASER.

Pierantoni, U.

'og L'origine di alcuni organi della Icerya purchasi e la simbiosi ereditaria.

Bull. Soc. Napoli, Vol. 23, pp. 147-150.
'ioa Origine e struttura del corpo ovale del Dacty'.opius citri e del corpo

verde del Aphis brassies:. Boll. Soc. Nat. Napoli, Vol. 24, pp. 1-4.
'iob Ulteriore osservazione sulla simbiosi ereditaria degli Omotteri. Zool.

Anz.. Vol. 36, pp. 96-111, Figs. 10.
'12 Studi sullo sviluppo d'Icerya pnrchasi Mask. Part I. Origine ed evolu-

zione degli element! sessuali femminili. Arch. Zool. Ital., Vol. 5, pp.
'133 Studi sullo sviluppo d'Icerya pnrchasi Mask. Part II. Origine ed evo-

luzione degli organi sessuali maschili. Arch. Zool. Ital., Vol. 7, pp.

243-274, Tav. 3.
'i3b Struttura ed evoluzione dell'Organo simbiotico di Pseudococcus citri

Risso e ciclo biologico del Coccidomyces dactylopii Buchner. Arch.

f. Protistenk., Vol. 31, pp. 300-316, Taf. 3.
Putnam, J. D.

'80 Biological and Other Notes on Coccidse. Proc. Davenport Acad. Sci.,

Vol. 2, pt. 2, pp. 293-347, I pi.

Sanders, J. G.

'05 The Cottony Maple Scale. Circ. Bur. Entom., U. S. Dept., n. s., No.

64, pp. 6.
Shinji, G. 0.

'19 Embryology of Coccids, with Special Reference to the Ovary, Origin
and Differentiation of the Germ Cells, Germ Layers. Rudiments of
the Mid-gut, and the Intracellular Symbiotic Organism. Journ.
Morphol., Vol. 33, pp. 73-167, Pis. 20.
Sulc, K.

'07 Kerminicola kermesina n. gen., n. sp., und physbkermina n. sp., neue
Mikroendosymbiotiker der Cocciden. SB. kgl. bohm. Gesellsch. Wiss.
Prag., Jahrg. 1906, art. 19, pp. 6, Figs. 2.

'n Pseudovitellus und ahnliche Gewebe der Homopteren sind wohnstatten
symbiotischen Saccharomyceten. SB. kgl. Bohm. Gessellsch. Wiss.
Prag. Jahrg. 1910, art. 3, pp. 39, Figs. 18.
Teodoro, G.

'12 Richerche -sull'emolinfa dei lecanini. Atti Acad. ven.-trent.-istr., Anno

5, pp. 72-84.

'16 Osservazione sulla ecologia delle cocciniglie con speciale riguardo alia
morfologia e alia fisiologia de questi insetti. Redia, Vol. 11, pp.
129-209, Figs. 3, Pis. 3.
'18 Alcune osservazione sui saccaromiceti del Lecanium persiccc Fab. Re

dia. Vol. 13, pp. 1-5.
Vejdovsky, F.

'07 Bemerkungen zum Aufsatze des Herrn Dr. H. Sulc ueber Kerminicola
kermesina. SB. kgl. bohm. Gesellsch. Wiss. Prag, Jahrg. 1906, art.
pp. 6-12, Fig. i.

32O CHARLES T. CRUES AND RUDOLPH W. CLASER.

EXPLANATION OF PLATES.
PLATE I.

1. Cross-section of overwintering Coccid, showing distribution of symbionts
in the fat-body. Low magnification.

2. Portion of fat-body at higher magnification with symbionts included in
adipose cells. Magnification about 400 diameters.

3. Similar portion, showing more abundant symbionts. Magnification about
400 diameters.

4. Symbionts in adipose tissue which is apparently in prccess of disinte-
gration. Magnification about 400 diameters.

5. Smear from culture of symbionts in liquid medium after 48 hours incu-
bation. Magnification about 400 diameters.

BIOLOGICAL BULLETIN, VOL XL.

C. T. BRUES AND R. W. GL'SER.

322 CHARLES T. BRUES AND RUDOLPH W. GLASER.

PLATE II.

6. Form of young colonies of symbiont after 72 hours incubation on potato
agar plate, viewed by reflected light. Magnified 5 diameters.

7. Group of colonies of symbiont on nutrient agar after 10 days incuba-
tion. Magnified 4 diameters. Note small size of colonies and paucity of
processes.

8. Group of colonies on the plate illustrated in figure 6, at same magnifi-
cation, viewed by transmitted light, to show internal structure and manner in
which radial processes develop.

9. Development of mycelia in liquid culture of symbiont after prolonged
incubation of 10 days. Magnified about 60 diameters.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE II

C T. BRUES AND R. W GLASER.

324 CHARLES T. BRUES AND RUDOLPH W. GLASER.

PLATE III.

10. Gross appearance of colonies of symbiont after prolonged incubation
(12 days) on potato agar. Viewed by reflected light to show peripheral proc-
esses. Magnified 3 diameters.

11. Portion of same plate at same magnification, viewed by transmitted
light, to show internal mycelial structure.

BIOLOGICAL BULLETIN, VOL XL.

PI AIE III.

C 1 BRUES AND R. W. GLASER.

THE MIGRATION OF THE PRIMARY SEX-CELLS OF
FUNDULUS HETEROCLITUS.

A. RICHARDS AND JAMES T. THOMPSON,
ZOOLOGICAL LABORATORY, UNIVERSITY OF OKLAHOMA.

The origin of the primary sex-cells in vertebrates is a problem
which has received considerable attention during late years. Ex-
tensive summaries of the literature upon this subject may be
found in the articles of Allen, 1911, and Jordan, 1917. Since no
complete agreement with regard to details has as yet been reached,
it is perhaps desirable to review briefly the earlier investigations,
and to point out any discrepancies in the results already obtained
which would seem to require further study. It would seem that
more evidence is necessary to warrant safe conclusions on the
matter.

Waldeyer (18/0) first described the differentiation of the sex-
cells from the " germinal epithelium " of a four-day chick. His
view of sex-cell origin from the mesothelium covering the meso-
nephros was accepted at the time, and has been supported even by
recent investigators. In 1880 Nussbaum advanced a rival theory
as a result of his observations on the embryology of the trout and
frog. He held that the sex-cells were of blastomeric origin, and
further that there was an extra-regional segregation and a migra-
tion to the germ gland. Weismann (1886) popularized this idea
in his work on the " continuity of the germ plasm."

Since the time of Nussbaum the evidence against the " germinal
epithelium" idea has steadily increased. A number of investi-
gators (Hoffman, 1892; Eigenmann, 1892; Beard, 1900; Woods,
1902; Allen, 1906, 1907, 1911; Dodds, 1910; Swift, 1914, 1915,
1916; Jordan, 1917) have failed to find any conditions not in ac-
cord with Nussbaum's theory.

However other recent workers (Firket, 1914, 1920; von Beren-
berg-Gossler, 1914) have been unable to accept this interpretation
of the activities of these cells. According to their viewpoint, the

325

326 A. RICHARDS AND JAMES T. THOMPSON.

migration of the primary sex-cells is reduced to a mere phylo-
genetic vestige and is without any great genetic significance.
Firket speaks of the primordial germ-cells as " primary genital
cells " which, after migration disintegrate in the germ gland,
being replaced by the true of " secondary genital cells " which
arise from the peritoneal cells of the germ gland. Von Beren-
berg-Gossler regards 'them as mesodermal wandering cells of late
endodermal origin, and describes them as contributory in the for-
mation of the Wolffian ducts.

In our study of this general problem in Fundiilus a number of
questions have arisen as separate phases of the matter. The blas-
tomeric origin of the sex-cells, their path and method of migra-
tion, and their history after reaching the germ gland are all mat-
ters requiring separate study. This paper has, as its special aim,
the definite identification of the primary sex-cells and the deter-
mination of the germinal path in Fundiilus embryos ; that is, it
is concerned with the second question listed. As yet our data
upon the first question is inadequate and we have not enough
material for a study of the third.

MATERIAL AND METHODS.

The material for this investigation consisted of the eggs of the
teleost, Fundiilus hctcroclitus and was collected at Woods Hole
in the summer of 1919. Care was exercised to insure an approxr
mately uniform fertilization of the ova by mixing them with
chopped testis. Two extensive series w^ere preserved during the
summer. Although accurate records were kept as to the age of
each group, they are of only jiominal value in this investigation,
since environmental and individual differences cause variations in
development of embryos of like age.

All embryos in these two series were fixed in Bouin's fluid and
stained by the familiar " long method " for iron haematoxylin.
Other material in various fixatives was also available for com-
parison. No trouble was experienced in obtaining slides which
show clearly the cytological characteristics throughout the series
as far as described. A majority of sections were cut 4 micra thick,
but some were cut 5, 6 and / micra. The thickness of all sections

MIGRATION OF SEX-CELLS OF FUNDULUS HETEROCLITUS. 327

was of course recorded. Most of the observations were made
from serial transverse sections because they show the dorso-
ventral position of the sex-cells more clearly in relation to the
outstanding features of the developing embryo than do those cut
longitudinally.

OBSERVATIONS.
Criteria of the Primary Sc.r-cclls.

The enumeration of criteria for any group of cells as distin-
guished from all others in a series of embryos is a task which
promises but doubtful results. There can be no question however
that the primary sex-cells do have distinctive characteristics which
make them easily recognizable, during the resting stages, to one
who has had them under observation. It is not always feasible
positively to identify the cells during division.

Throughout the migration period these cells maintain the same
general characteristics. There are, to be sure, slight variations in
the ratio between the nuclear and cytoplasmic elements, in size
and in the character and arrangement of the chromatin granules ;
but these features may be observed only on close inspection,
rather than in a preliminary study of the primary sex-cells.

The primary sex-cells vary in diameter from 9 to 128 micra.
As contrasted with other cells they are spherical or ovoid with
very definite cell outlines. The nuclei conform to the general
shape of the cell body within which they are located. The cyto-
plasmic content is always clearer and takes less stain than that of
the surrounding cells. Likewise the achromatin of the nuclei is
quite clear, allowing the chromatic granules to stand out in bold
contrast. The limn network is directly beneath the nuclear mem-
brane, and due to this arrangement the chromatin granules are
distributed peripherally over the nucleus. This peripheral ar-
rangement of the chromatin is a constant distinguishing charac-
teristic not to be mistaken, for it is never produced in any other
cells. The linin network is connected to one, or more frequently
to two nucleoli which are located near the center of the nucleus.
No peculiar invagination of the nuclear membrane, such as was
reported by Dodds (1910), was observed in Fundnhis. An un-
usually large centrosome is, as a rule plainly visible in the cyto-

328 A. RICHARDS AND JAMES T. THOMPSON.

plasm. In older embryos these cells may be recognized by their
size, since they are larger than any others which may occur in the
same region.

Figures i, 2, 3 and 3a are surface views of typical sex-cells.
The peripheral arrangement of the chromatin has been empha-
sized in drawing Fig. <\b, by focusing upon a level with the center
of the nucleus. Fig. 40 was obtained by focusing higher on the
surface of the same nucleus. Thus Fig. 46 represents the chro-
matin knots in an optical section ; wrhile Fig. 40 shows them in a
surface view. If Fig. 40 were superimposed upon Fig. 4^ the
resulting composite \vould be a cell not unlike that represented in
Fig. 3, except that in the latter the knots have taken the familiar
granular appearance.

A positive identification of the primary sex-cells was first made
in a 24-day embryo. From this stage their path was followed
backwards, through all the intermediate phases of migration,
until they were no longer evident. It is considered expedient to
describe their position in the 24-day stage, so that no question
shall arise later as to the exact nature of the migrating cells
whose course is to be traced.

Having established the identification of the sex-cells in the late
embryo (24 days) their migration may be traced from their
earliest appearance up to this stage. Although this sequence is
contrary to our experimental procedure, it is believed to be more
easily followed by the reader.

zj-Day Embryo. 5.75 Mm. Long, — At the 24-day stage the
sex-cells lie in the sac-like anlagen of the germ glands, which
have formed dorsally and slightly laterally to the hind gut. Here
they are unquestionably recognizable (Fig. 5). These cells are
numerically inferior to the peritoneal cells which surround them
and which are beginning to take a very active part in the forma-
tion of the future sex gland. The size and position of the germ
gland anlagen in relation to the embryo is shown in Figs. 6 and 7.
No attempt has been made to ascertain the average number of
sex-cells which are present during this stage.

Whether these are the true sex-cells as maintained by many
investigators, or whether they later disintegrate and become re-
placed by "secondary genital cells" as indicated by Firket (1914,

MIGRATION OF SEX-CELLS OF FUXDULUf, HETEROCLITUS. 329

1920) and others, is a question which may be omitted from the
present discussion.

Observations of the Germinal Path.

From many available embryos the following were selected for
consideration because they constitute representative stages, and
are essential to a clear understanding of the migration of the
primary sex-cells.

Embryos from 46 to 50 Hours. — Three embryos of this very
early stage, designated in our material as B' 2iX, B' 2iL and
B' 23 respectively were carefully studied. Others were available
but they were used merely as checks on the three which are
reported.

The position of the sex-cells at this stage is most striking.
There is a wide range of distribution in each embryo. The most
anterior of the sex-cells \vere invariably farther along the ger-
minal path than were the more posterior ones of the same embryo.
To demonstrate this fact Fig. 20 has been drawn ; it is an exact
outline diagram of B' 2iX, reconstructed by the most accurate
means possible. The lateral extent of the neural tube, of the
mesoderm and the positions of the sex-cells were determined by
measuring from the median line. The thickness 'of the sections
was known. These determinations were plotted on millimeter
paper and the outline filled in as indicated by the plotted guides.
The exact antero-posterior positions of the sex-cells were deter-
mined by counting the sections of the serially sectioned embryo.
Figs. 9, 10 and 11 are outline drawings from the embryo B' 2iX
which was sectioned transversely, and Fig. 12 from B' 2iL which
was sectioned longitudinally. These drawings show the exact
positions of the sex-cells more clearly than would be possible in
a written description.

The letters A, B and C on Fig. 20 indicate the positions of the
sex-cells shown in Figs, n, 10 and 9 respectively. Fig. 11 illus-
trates clearly the position of the primary sex-cell in the extra-
embryonic region at the posterior of the embryo. Four sex-cells
are shown in Fig. 12 as being lateral to the undifferentiated endo-
dermal cell mass and ventral to the elongating tail. In Figs. 9
and 10 the migration has progressed proportionally to the devel-

330

A. RICHARDS AND JAMES T. THOMPSON.

D

TEXT FIG. A. Semidiagrammatic transection through the posterior of the
5o-hour embryo B' 23. The spots indicate the positions in which the greater
portion of the sex-cells are found at this stage. The rectangle includes the
area which is drawn in detail in Fig. 8. X 225.

•TEXT FIG. B. Transection from the iO5-hour embryo, showing the first
decided advance over the stage illustrated in Text Fig. A. Here the gut,
Wolffiau ducts and the ccelome have taken form. The rectangle includes the
area drawn in Figs. 14, 15 and 16. X 225.

TEXT FIG. C. Showing progress of development after (5 days. Figs. 17
and 18 are detailed drawings of the area included in the rectangle. X 225.

TEXT FIG. D. Transection through the developing gonads of the g-day
embryo, B 34. The sex-cells are collected ventral to the Wolffian ducts. The
dorsal mesentery shows a decided change from conditions found in earlier
stages. X 90.

TEXT FIG. E. From the 1 3-day embryo B 42, showing the effect of the
developing swim bladder. X 90.

MIGRATION OF SEX-CELLS OF FUNDULUS HETEROCLITUS. 33!

opment of the embryo at the regions represented. The sex-cells
lie between the periblast and the endoderm in Fig. 10 ; while in
Fig. 9 their position is below the mesoderm and lateral to the de-
veloping hind gut.

The positions of the most anterior sex-cells in embryo B' 23 are
indicated in Text Fig. A. The rectangle in this text figure in-
cludes the region which is drawn in detail in Fig. 8. Here a pri-
mary sex-cell is shown which is entirely free from any possible
connection with the lateral mesoderm. It can scarcely be said to
lie in, but rather lateral to the gut endoderm. It is half buried in
the periblast. This fact suggests intimate relation with this nu-
tritive layer. The cell figured is one of the few ever found with
an irregular outline. This might seem to suggest amoeboid activ-
ity, but this type is so extremely rare that it may be neglected
from consideration.

Fig. 13 from B' 23 shows a sex-cell which is .06 mm. to the rear
of the one just mentioned. It is plainly in that portion of the
lateral mesoderm which will develop into the splanchnic layer
upon the formation of the ccelome (about the third day).

Observations of these early embryos show several important
facts. The primary sex-cells are as truly characteristic and as
easily recognizable as any found in the germ glands of later stages.
They are located in the posterior half of the embryo, becoming
gradually more numerous as the anterior part of this region is
approached. Laterally they range from the extra-embryonic
region to within the lateral mesoderm and the edge of the de-
veloping gut. In general their progress along the germinal path
is directly proportional to the development of the embryo.

lOj-Hour Embryo. — Text Fig. B shows the relative positions
of the sex-cells in the 105-hour embryo, B' 26. As in previous
cases the rectangle indicates the area from which Figs. 14, 15 and
16 were drawn. These three figures from the same embryo illus-
trate the full extent of the migration at this stage. On the left
side of the embryo the sex-cells are found scattered all along the
splanchnic mesoderm, from the region very near the split in the
lateral mesoderm (Fig. 14) to that at the side of the gut (Fig.
16). On the right of Text Fig. B the sex-cells on the opposite
side of the embryo are shown massed lateral to the gut. Should

332 A. RICHARDS AND JAMES T. THOMPSON.

the lateral mesoderm fuse above the gut, the formation of the
dorsal mesentery would result and the position of the sex-cells
would be identical to that found in later stages. Ex. Figs. 17
and 18.

6-Day Embryo. 2.6 Mm. Long. — Text Fig. C represents the
position of the sex-cells as found in the 6-day embryo. At this
time they are apparently in a state of rapid migration from the
loose mesenchyme dorsal to the hind gut, to the positions ventral
to the Wolffian ducts. Because of the laterally compressed con-
dition of the embryo, which was due to the softness of the paraffin
at the time of cutting, the transections are not exactly typical.
However this embryo has been used since it represents most
clearly the transitional stage between those figured in Text Fig.
B and D. Figs. 17 and 18 are detailed drawings of the 6-day
stage. They illustrate the complete extent of the migration in the
mesentery. The majority of the sex-cells were in the dorso-
ventral position indicated by the cells in Fig. 17, while only a few
were in that shown in Fig. 18. The more anterior sex-cells were
farther along in the germinal path (being nearer the Wolffian
ducts) than the more posterior ones. The position of the germ
gland anlagen ventral to the Wolffian ducts is illustrated in Text
Fig. D.

i^-Day Embryo. 4 Mm. Long. — Text Fig. E represents the
position of the germ gland anlagen as found in the 1 3-day embryo
B 42 (4 mm.). The rectangle in Text Fig. E includes the region
which is drawn in detail in Fig. 19. Rarely more than one sex-
cell is found at this stage in any one section of a germ gland
anlage. The anlagen are little more than protuberances from
the peritoneum, containing relatively few peritoneal cells (al-
though the sex-cells are surrounded by them) and they have not
yet reached the future position of the gonads. It is obvious that
the sex-cells which are contained in the peritoneal sac are all
pushed ventrally by the developing swim bladder. One cell was
observed in the position indicated by the cross in Text Fig. E. It
was not included in the peritoneal sac and seemed apparently
helpless in the loose mesenchyme ventral to the Wolffian duct.
This cell had been delayed in reaching this position, had not been
included in the sac, and in consequence of this fact it had not been

MIGRATION OF SEX-CELLS OF FUNDULUS HETEROCLITUS. 3^3

influenced by the action of the swim bladder. One of these lost
cells is shown in the mesentery in Text Fig. D. The future of
such cells is an open question.

A count of the sex-cells in this embryo, B 42, gave 64. There
was never any question of recognizing these cells, for no cells of
doubtful character were observed. Four of these cells found
were in the mesentery above the gut, and one was in the loose
tissue ventral to the Wolffian duct. No cases of disintegrating
sex-cells were observed in our Fnndnlus material, although such
conditions are reported by some investigators.

The 2^-day embryo shows the next advance in the germinal
path. This stage has been considered perviously in connection
with " Criteria of the Primary Sex-Cells."

TABLE OF AVERAGE DIAMETERS.

For the purposes of this investigation the diameters in micra
were found by averaging the long and short dimensions of the
cell and nucleus. By this method a number of representative sex-
cells of embryos in all stages of migration were measured under
the oil immersion and the following results were obtained.

Embryo.

Average.

i.B'2iX (embryonic region)
46 hour6!

Cell

IO.O

IO.S

o 6

10 4

IO I

2.B'2iX (extra-embryonic)
46 hours

Nucleus. .
Cell .

6.0

12.8

6.4

II. 2

5-6

II. 2

5-2

II. e

5-8
117

3.B'26 105 hours

Nucleus. .
Cell . . .

6.0

Q.Q

6.1

II. O

6.5
10.7

6-5
10.4

IO.2

6.3

IO A

4.B 30 6 days . .

Nucleus. .
Cell .

6.1

12. T,

6.2
12.4

6.2

12.7

6.1

II. 2

5.8

6.1

122

S.B 42 13 days

Nucleus. .
Cell

6.6
ii. =;

6.6
II. 5

7-0

I I.Q

6.2
IO.7

Q.O

1 1.<;

6.6

II. 2

6.B 65 24 days . .

Nucleus. .
Cell .

7.2
i i.s

6.6
11. 8

8.0

IO.2

7.8

IO A

5-7
0 O

6.3

9C

6.9

IO A

Nucleus. .

6.1

7.8

7-3

7.0

6.2

6.6

7.2

Multiplication of tlic Sc.v-cclls.

The early distribution of the sex-cells (Figs. 20 and 21) is best
explained, we believe, in connection with the streaming of the
organ-forming substances which contribute materials to the em-
bryo body. In most processes of this nature not only cell trans-

J34 A- RICHARDS AND JAMES T. THOMPSON.

portation but cell division takes part. Certain workers with other
forms have held that the movement of the cells into the anlagen
of the gonads is not the only factor responsible for their increase,
but that multiplication actually occurs during the period of trans-
location. Mitotic figures have never been observed in Fundulus
among the recognizable sex-cells which are within the embryo,
although a most thorough search has been made for them in many
embryos at all stages of development. A count of these cells in
several specimens in various stages reveals the fact that there is a
tendency for their number to vary more or less from the average
established (67). However there is not enough variation to con-
vince one that there is any marked multiplication of the sex-cells
during the migration period. These facts naturally lead to the
conclusion that the first period of multiplication takes place in
the extra-embryonic region.

In the description of the earliest embryos referred to in this
report and in the figures presented, emphasis has been placed
upon the fact that the primary sex-cells in any one embryo are
not in the same phase of migration. Furthermore observations
upon all stages show that development becomes more advanced
anteriorly than posteriorly. It is of further interest that in em-
bryos containing sex-cells both within and without the body, the
number falls below the average for older stages. These condi-
tions and the fact that no sex-cells have been found in any region
other than that already described, suggest the explanation that
these cells multiply in the extra-embryonic region. Indeed the
four sex-cells illustrated in Fig. 12 may indicate recent cell divi-
sion by their very association. If they are not of recent and iden-
tical origin they would probably be farther separated than they
are in this figure. These views are presented only tentatively,
due to lack of sufficient material to warrant definite statements
on this multiplication ; for no mitotic figures have ever been seen
in the extra-embryonic region to substantiate this belief in a divi-
sion as suggested. Our material has not permitted a careful study
of this matter. However considering the longitudinal distribu-
tion of the sex-cells in the earliest available embryos, and their
tendency to approach a common average number in each indi-
vidual, one is inclined to regard them as being of unquestionably

MIGRATION OF SEX-CELLS OF FUNDULUS HETEROCLITUS. 335

earlier origin than it has beeli possible thus far to trace them.
It seems not unreasonable to believe that the fore-runners of
these cells have been segregated at a time very early in the de-
velopment of the germ ring.

DISCUSSION AND CONCLUSIONS.

This paper attempts to identify definitely the sex-cells which
are present in the 24-day embryo as the " primordial germ cells "
of previous writers, or as the " primary genital cells " of Firket.
It also presents evidence on the manner in which these cells reach
their final destination.

The method of embryo formation in the teleosts has a bearing
upon the question of sex-cell migration in Fnndulus. It will be
recalled that the anterior portion of the embryo is formed from
the head fold, which may perhaps be nothing more than a thick-
ening on the germ ring ; while the body or posterior portion is to
be regarded as the result of the developmental process termed
concrescence. It is only this latter portion of the body that is
involved in the formation of the sex-cells. The eggs have a
large amount of yolk, and a very distinct germ ring. As cell
proliferation takes place, the germ ring moves gradually down-
ward over the yoke mass. The primitive streak moves back-
ward and receives the converging limbs of the germ ring
posteriorly. The material of the halves of the germ ring, after
fusion, is differentiated into the embryo posterior to the head
process. The rudiments of the embryo body are not clearly
marked out in Fnndulus until the germ ring is completely closed.

The earliest primary sex-cells which we have located are from
embryos in which the germ ring has been closed but a few hours,
and in which the tail is just beginning to elongate. Their posi-
tion in the extra-embryonic region lateral to the undifferen-
tiated endodermal cell mass at the posterior half of the embryo
is indicated in Fig. 20. In other embryos of the same stage of
development, numerous primary sex-cells are present in prac-
tically the identical relation to the embryo that is clearly demon-
strated in Fig. 20. These sex-cells invariably lie just above the
periblast and are associated with the sheet of cells which is a
lateral expansion of the undifferentiated endodermal cell mass

336 A. RICHARDS AND JAMES T. THOMPSON.

(peripheral endoderm, Allen). The complete germinal path
from this position to one lateral to the hind gut may be followed
in almost any embryo of from 46 to 50 hours. This very advan-
tageous condition is made possible by the greater development
near the middle of the embryo, for it is only a natural result of
embryo formation by concrescence that development is progres-
sively greater anteriorly from the point of convergence of the
germ ring.

These cells are transported from the edge of the embryonic
region medially, to positions just beneath or within the endoder-
mal cell mass, as the case may be. They are carried passively
from one position to another by the same forces of growth which
bring together the halves of the germ ring. The influence of
this factor can scarcely be over emphasized. Although not out-
wardly as apparent as in earlier stages, these forces are neverthe-
less responsible for the flowing of the streams of embryonic
material towards the future position of the organs which are to
develop therefrom.

The sex-cells come to lie within these shifting layers of em-
bryonic endoderm and mesoderm and naturally accompany these
layers in their changes of position. Because of the fact that the
movement of the sex-cells is not active but rather dependent
upon that of surrounding layers, the expression " migration "
seems rather unfortunate. Some term such as " translocation "
would perhaps be more truly expressive of the actual conditions.

These cells come to lie in the edge of the embryonic region,
and when a portion of the undifferentiated cell mass gives rise to
gut endoderm and another to lateral mesoderm, they follow one
layer or the other. The sex-cells follow one or the other of these
layers until they reach a position lateral to the newly formed gut.
Which layer is chosen apparently depends upon chance. Those
cells which have been carried in the edge of the endoderm never
enter the gut, but move dorsally from the side of it into the lat-
eral mesoderm. Here they join the sex-cells which have been
carried in the mesoderm. By this time the split, resulting in the
formation of the ccelome between the splanchnic and somatic
mesoderm has taken place. Although the sex-cells are asso-
ciated with all parts of the lateral mesoderm before the forma-

MIGRATION OF SEX-CELLS OF FUNDULUS HETEROCLITUS.

tion of the coelome, it is a noteworthy fact that they never occur
within the somatic layer after differentiation.

From this position lateral to the hind gut the cells are in the
general dorsal movement of the mesoderm which eventually re-
sults in the formation of the intestinal mesentery. The cells
from either half of the embryo remain apart and seem to lie in
separate streams of mesodermal cells which are flowing toward
the Wolffian ducts. But although there may be a pause here, at
no time do the sex-cells appear to establish any intimate relation
with the cells of these ducts (Text Fig. D). From the evidence
at hand, an explanation of the function of these cells which
makes them contributory to the development of the already well-
formed Wolffian ducts, as suggested by certain investigators, does
not seem plausible in Fnndnlns.

As the sex-cells reach a position nearly ventral to the Wolffian
ducts they become surrounded by a single layer of peritoneal
cells. This covering develops until the position of the future sex
organs is attained ; the sex-cells then rest in sac-like protuber-
ances from the peritoneum, the germ gland anlagen. Assisting
in the movement which brings the sex-cells into their future posi-
tions, are several factors entirely external to the germ glands.
For example, there is a rapid proliferation of the loose mesen-
chyme dorsal to the gut and the development of the swim blad-
der which results in a median down pushing. The ventral
movement (Text Figs. D and £) from the region of the Wolf-
fian ducts is clearly due to the wedge-like effect produced by the
growing swim bladder. That this process is necessarily passive
is evident from the fact that amoeboid activity of the cells in-
cluded within the germ gland would be unable to produce any
change in its position.

From the evidence in Fundulus it is apparent that the sex-cells
enter the embryo and are located in the germ glands by the same
forces that are influential in the distribution of the other organ
forming substances of the body. Their " migration " is not to be
looked upon as different from that of any other group of cells.
But while the sex-cells are not amoeboid, there is nevertheless
reason for misunderstandings which have arisen regarding their
activities. In the first place they are relatively few in compari-

33^ A. RICHARDS AND JAMES T. THOMPSON.

son to the great numbers of cells in the surrounding tissues. Al-
though the entire mass of cells is continuously in motion, only the
movement of the sex-cells is at all noticeable. They are shifted
about by the active surrounding tissues and naturally assume
slightly irregular outlines at times, due to the unequal tension
upon the cell membrane. Through a misinterpretation of the
conditions within the embryo these sex-cells may easily be ac-
credited with peculiar powers of locomotion. A sex-cell, as a
slowly drifting cloud, can be seen gradually to change its posi-
tion; but movements of the tissue cells about it, due to their loca-
tion in continuous layers, are so inconspicuous as to go unnoticed.
Because of this, the movement of the sex-cells should be consid-
ered merely as the passive indication of the rate and direction
of progress of contiguous layers.

SUMMARY.

1. The earliest primary sex-cells found in Finidnliis were
located in the peripheral endoderm, lateral to the posterior half
of the 46-hour embryo. No sex-cells were observed in that part
of the embryo which develops from the head fold.

2. The germinal path leads from the peripheral endoderm, into
the border of the undifferentiated endodermal cell mass. When
this cell mass splits to form gut endoderm and lateral mesoderm,
the sex-cells proceed medially with either layer. By the time the
gut is formed, these cells are lateral to it ; they all eventually be-
come located in the splanchnic mesoderm of this region. From
here the sex-cells migrate dorsal to the hind gut, thence to the
region ventral to the Wolffian ducts. Here they become sur-
rounded by peritoneal cells which form the somatic portion of
the gonads. From this position the germ gland anlagen are
shifted back to their final location dorsal to the gut.*

3. There is very little multiplication of the sex-cells during the
period of migration. Division apparently takes place in the
extra-embryonic area, and is not renewed to any marked extent
until after the sex-cells become located in the germ glands.

4. The constant distinguishing characteristics insure positive
identification of these cells throughout all phases of their migra-
tion, and leave no reason to question their identity as being the
"primordial germ cells" of previous writers.

MIGRATION OF SEX-CELLS OF FUNDULUS HETEROCLITUS. 339

5. Migration is passive, being due to forces of growth which
are altogether external to the cells themselves. These forces of
growth are factors common to the development of the organs
formed in the body of the teleost embryo.

6. Evidence derived from this study of Fundulus is an abso-
lute harmony with the theory of early segregation of these pri-
mary sex-cells.

NOTE

Some time after the manuscript of this paper had been sent
to the press, an extensive article by Okkelberg, entitled, " The
Early History of the Germ Cells in the Brook Lamprey, Ento-
sphcnus zi'ildcri (Gage), up to and Including the Period of Sex
Differentiation," appeared (Jour. Morph., Vol. 35, No. I, 1921).
This article contains much data and many important conclusions,
and it is to be noted (on pages 35 and 36) that the author, in
considering the sex cells, has discussed their methods of migra-
tion. It is of great interest that the conclusions reached by Ok-
kelberg on this matter for the lamprey are very similar to our
own upon Fnndnlus.

BIBLIOGRAPHY.
Allen, B. M.

'04 The Embryonic Development of the Ovary and Testis of the Mam-
malia. Amer. Jour, of Anat., Vol. 3.

'n The Origin of the Sex-Cells of Aiuia and Lepidostcns. Jour, of Mor-
phology, Vol. 22, No. i.
Dodds, G. S.

'10 Segregation of the Germ-Cells of the Teleost Lopluits. Jour, of Mor-
phology, Vol. 21.
Eigenmann, C.

'92 On the Precocious Segregation of the Sex-Cells of Cymatogaster.

Jour, of Morphology, Vol. 5.
'96 Sex Differentiation in the Viviparous Teleost Cymatogaster. Arch.

fur Entw. Mech., Bd. 4.
'03 Viviparous Fishes of the Pacific Coast (Cymatogaster aggregates}.

Bull. U. S. Fish Comm., Vol. XII, p. 412.
Firket, F.

'20 On the Origin of the Germ-Cells in the Higher Vertebrates. Anat.

Record, Vol. 18, No. 3.
Jarvis, May.

'08 The Segregation of the Germ Cells of Phrynosoma cornatitiu. Pre-
liminary Note. BIOL. BULLETIN, Vol. 15, No. 3.
Jordan, H. E.

'17 Embryonic History of the Germ-Cells of the Loggerhead Turtle. Pub.
No. 251 Carnegie Inst. of Wash., pages 313 to 344.

34O A. RICHARDS AND JAMES T. THOMPSON.

King, H. D.

'08 The Oogenesis of Bufo Lentiginosus. Jcur. of Morphology, Vol. 19

P- 369.
Nussbaum, M.

'80 Zur Differenzierung des Geschlechts im Tierreich. Arch. f. mikr. Anat.,

Bd. 18.
Swift, C. H.

'14 Origin and Early History of the Primordial Germ-Cells in the Chick.

Amer. Jour, of Anatomy, Vol. 15, No. 4.

'15 Origin of the Definitive Sex-Cells in the Female Chick and their Rela-
tion to the Primordial Germ-Cells. Amer. Jour. Anat., Vol. 1 8, p. 441.
'16 Origin of the Sex-Cords and Definitive Spermatogonia in the Male

Chick. Amer. Jour. Anatomy, Vol. 20, p. 375.
Wilson, H. V.

'89 Embryology of the Sea Bass. Bull. U. S. Fish Comm., Vol. 9.
Woods, F. A.

'02 Origin and Migration of the Germ-Cells in Acanthias. Amer. Jour
Anat., Vol. i, No. 3.

EXPLANATION OF ILLUSTRATIONS. '

All figures in this report were drawn with the aid of a camera lucida.
Any lens combinations which were considered necessary to produce the best
results were used. The magnification as given for each figure was calculated
carefully and is correct for the reproductions as they appear on these plates.

ABBREVIATIONS.

a., anus. Mes., gut mesentery.

Ao., aorta. Af, notochord.

B., blood cells. Nt., neural tube.

c., centrosome. nu., nucleolus.

cr., chromatin knots. P., peritoneum.

Co., ccelome. pe., periblast.

Ect., ectoderm. P.N., periblast nucleus.

EM., endodermal cell mass. 5"., sex-cell.

En., gut endoderm. sw., swim bladder.

G., germ gland anlagen. So., somatic mesoderm.

g.t gut. Sp., splanchnic mesoderm

L.M., lateral mesoderm. T., elongating tail.

li., linin network. Wo., Wolffian duct.

M., mesoderm.

342 A. RICHARDS AND JAMES T. THOMPSON.

DESCRIPTION OF ILLUSTRATIONS.

PLATE I.

FIG. i. A typical primary sex-cell from a los-hour embryo (B' 26).
Showing the large centrosome, two nuck-oli and chromatin knots scattered
oVer the periphery of the nucleus. X 1420.

FIG. 2. A typical sex-cell from a 6-day embryo (B 30). Chromatin gran-
ules finer than in the preceding cell. X 1420.

FIG. 3. Sex-cell from a loo-hour embryo (B' 25-3). From the extra-
embryonic region. It is closely associated with the periblast. In other sec-
tions the peripheral endoderm may be seen out over this area. X 1420.

FIG. 3«. Sex-cell from a g-day embryo (B 34). Chromatin arrangement
is intermediate between that found in Figs, i and 2. X 1420.

FIG. 40. Surface view of a nucleus from a sex-cell in a io5-hour embryo,
showing the chromatin knots in surface view. X 1420.

FIG. 4&. From the same nucleus as the one used in Fig. 40. Obtained by
focusing upon the center of the nucleus, illustrating the chromatin knots in
optical section. X 1420.

FIG. 5. Showing the sex-cells in the germ gland anlagen dorsal to the
hind gut of a 24-day embryo. The function of the peritoneal cells at this
stage is quite evident in this illustration. The sex-cells are completely sur-
rounded by mesoderm. X 700.

FIG. 6. Transection through a 24-day embryo (B 65-2) taken at the posi-
tion indicated by the plane X-X' in Fig. 7. Showing the position of the
gonads in relation to other parts of the embryo ; especially as regards the
developing swim bladder. X 120.

FIG. 7. Longitudinal section lateral to the median line of the 24-day em-
bryo (B 65). Illustrating the position of the gonads as being the same as in
the adult. Migration ceases at this point. The activities within the gonads
after they have reached this point of development will not be considered at
this time. X 25.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE I.

4b

A. RICHARD1 AND J. T. THOMPSON

344 A- RICHARDS AND JAMES T. THOMPSON.

PLATE II.

FIG. 8. A sex-cell from the so-hour embryo (B' 23). Showing in detail
the region at the edge of the embryo, where the germ layers meet the peri-
blast. The sex-cell is in the edge of the gut endoderm, entirely removed
from any connection with the lateral mesoderm, and is partially imbedded in
the periblast. Several very early cells have been found in this relation to the
periblast, but so far it has been impossible to establish any significance to this
fact. Due to exertion of unequal tension upon the cell membrane, the cell
outline appears slightly irregular. X 1400.

FIG. 9. Semi-diagrammatic transection of the 46-hour embryo B' 21 X
taken at the position indicated by the line " C " in Fig. 20. At this early
stage the lumen of the gut is not formed completely, even in this most ante-
rior region. The sex-cells lie between the periblast and the lateral mesoderm,
at the side of the developing hind gut. X 300. (The positions of the sex-
cells in the germinal path correspond to certain stages in development of
the gut.)

FIG. 10. Transection of embryo B' 21 X taken at the position indicated
by the line B in Fig. 20. Here the gut endoderm and the lateral mesoderm
are differentiated to a certain extent, although the splitting of the endodermal
cell mass has not yet occurred. The sex-cell lies above the periblast in the
edge of the cell mass. X 300.

FIG. ii. Transection of embryo B' 21 X taken at the line "A " in Fig. 20.
The sex-cell illustrated is in the extra-embryonic region, associated closely
with the peripheral endoderm. X 300.

FIG. 12. A longitudinal section through the elongating tail of the 46-hour
embryo B' 21 L. The 4 sex-cells illustrated are in the peripheral endoderm
at the extreme posterior of the embryonic area. X 300.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE II.

10

A. RICHARDS AND J. T. THOMPSON.

346 A. RICHARDS AND JAMES T. THOMPSON.

PLATE III.

FIG. 13. Sex-cell from the so-hour embryo B' 23. It lies in the extreme
edge of the lateral mesoderm. just dorsal to its separation from the periblast.
This cell lies in that portion of the mesoderm which will give rise to the
splanchnic layer. X 700.

FIG. 14. A sex-cell in the splanchnic mesoderm of the io5-hour embryo
B' 26. The cell is just medial to the point of differentiation between the
somatic and splanchnic layers. This stage also shows an advance over the
one illustrated in Fig. 13, in that the gut and Wolman duct have taken form.
X 700.

FIG. 15. A group of sex-cells from the region anterior to that drawn in
Fig. 14. Here the sex-cells are approaching the side of the hind gut. X 700.

FIG. 1 6. A sex-cell from embryo B' 26, in the region anterior to those
illustrated in Figs. 14 and 15. The more medially placed cell is in the
splanchnic mesoderm lateral to the gut. As this layer grows up over the gut
to form the dorsal mesentery, the sex-cells will naturally be brought to lie in
this region. X 700.

FIG. 17. Four sex-cells from the 6-day embryo B 30 in the mesentery
dorsal to the gut. This shows in detail the region included in the rectangle
in Text Fig. C. The extraordinary width of the mesentery at this stage is
doubtless due in part to the presence of the sex-cells. X 420.

FIG. 1 8. This figure illustrates the extremes of the migration at this stage.
No cell was found at any earlier stage than the one near the gut, none later
than that ventral to the Wolffian duct. From embryo B 30. X 420.

FIG. 19. Showing the development of the gonads in the 1 3-day embryo
B 42. The sex-cells are fixed in sac-like protuberances from the peritoneum.
From this figure it is possible to obtain an idea of the effect produced by the
rapid growth of .the swim bladder, in literally pushing the gut and all related
tissues ventrally. It is also interesting to observe that the peritoneal cover-
ing renders this cell dependent upon surrounding tissues for movement to
its final position dorsal to the gut. X 420.

BIOLOGICAL BULLETIN, VOL. X'

PLATE III.

A. RICHARDS AND J. T. THOMPSON.

34-8 A. RICHARDS AND JAMES T. THOMPSON.

PLATE IV.

FIG. 20. A diagrammatic reproduction of embryo B' 21 (46 hours) dem-
onstrating the distribution of the sex-cells at this early stage. The variation
in anterior-posterior development is very noticeable and is explained fully in
the text. X 65.

FIG. 21. Reproduction of embryo B' 26 (105 hours) constructed simi-
larly to Fig. 20. Showing the positions of the sex-cells as being more medially
placed than in the earlier stage. The cells on the right are not as far along
with migration as those on the left ; the former group migrated with the
mesoderm and the latter followed the gut endoderm. X 65.

BIOLOGICAL BULLETIN, VOL XL.

PLATE IV

7.1

A. RICHARDS AND J. T. THOMPSON.

SPERMATOGENESIS OF APHIDS ; THE FATE OF THE
SMALLER SECONDARY SPERMATOCYTE.

H. HONDA.

CONTENTS.

I. Introduction ^49

II. Method 35o

III. Stomaphis yanois 350

1 . Primary Spermatocyte 350

2. Larger Secondary Spermatocyte 351

3. Smaller Secondary Spermatocyte 352

4. Smaller spermatid 353

IV. Neothomasia populicola and Macrosiphum ambrosia: 356

V. Review 358

VI. Summary 360

I. INTRODUCTION.

It has been shown by Morgan and von Baehr that in aphids
the primary Spermatocyte divides unequally producing larger
and smaller secondary spermatocytes. The larger secondary
Spermatocyte undergoes a second maturation division, and pro-
duces two equal-sized spermatids which transform into func-
tional spermatozoa. The smaller secondary Spermatocyte, which
has received fewer chromosomes is said to degenerate. Only two
similar spermatozoa, consequently, are formed from a primary
spermatocyte.

Von Baehr (1909 and 1912) states that he very rarely observed
the development of the smaller secondary spermatocyte to the
prophase of the second maturation division, but that it does not
divide. Stevens (1909) says that in a preparation which has
unfortunately been lost, the anaphase of the smaller secondary
spermatocyte was seen, and that such stage may also be distin-
guished among the degenerating spermatocytes. Morgan (1915)
states : " The small cell is left with two chromosomes and a small
amount of cytoplasm. It never divides again, and later degen-
erates. Stevens was inclined to think that the small cell may

349

35O H. HONDA.

sometimes show a division figure, which subsequently fades away,
but I have never seen a case of this kind."

In Macrosiphum ambrosia: and Ncothomasia popnlicola I have
observed the late telophase of the smaller secondary spermato-
cytes. In Stomaphis yanois, moreover, I have found that the
smaller secondary spermatocytes divide and form equal-sized
spermatids which are much smaller than the larger spermatids.
These smaller spermatids develop and reach the sustentacular
cells with the developed larger spermatids ; they, however, fail to
attach to the sustentacular cells. Thus their development ceases ;
they, therefore, do not fully transform into spermatozoa, but
retrogress and form spherical cells, which attach themselves to
the epithelium of the cysts of the testes. A further account of
this will appear in the following pages.

The work on the spermatogenesis of Stomaphis yanois was
done in the Tokyo Higher Normal College, and the work on the
other aphids has been done in the University of Chciago. The
writer's thanks are due to Prof. F. R. Lillie and Prof. S. Yama-
nouchi, who gave him many suggestions and much help. The
writer also wishes to thank Prof. A. Oka and Prof. U. Takakura
for their kindness during his stay in Tokyo. For the identifica-
tion of the aphids the writer is indebted to Dr. A. L. Quaintance
and Dr. A. C. Baker.

II. METHODS.

Both the males and parthenogenetic females were fixed in
either strong Flemming's, Zenker's or a mixture of absolute alco-
hol one part, acetic acid one part, and saturated aqueous solution
of corrosive sublimate two parts. Sections were cut 3, 5 and 10
micra in thickness ; most of them, however, were cut 5 micra
thick. They were stained with Heidenhain's iron-hematoxylin
followed by eosin or borax carmin.

III. STOMAPHIS YANOIS.
i. Primary Spcrmatocytc.

Figs, i and 2 show the primary spermatocyte prophase. In
Fig. i, one of the chromosomes is formed, and in Fig. 2, the proc-

SPERMATOGENESIS OF APHIDS. 35!

ess of the formation of the chromosomes is almost finished.
There are ten chromosomes, five larger and five smaller, in the
equatorial plate of the first spermatocyte division, and they are
connected with one another by linin threads as is shown in Figs.
4 and 5. The side view of the mitotic figure shows centrosomes
of about the same size agreeing with von Baehr's observation on
Aphis saliceti (Fig. 3).

In the anaphase unequal cell division is indicated. The larger
and smaller daughter cells are connected by a bridge of cyto-
plasm, and elongated lagging chromosomes lie between the chro-
mosomes passing to the daughter cells (Fig. 7). The lagging
chromosomes do not show any tendency to go to the larger cell
at this time, but after the nuclear membrane is formed, the lag-
ging chromosomes enter the larger cell (Fig. 8). It is interesting
to note that the size of the nuclear membrane -is larger in the
larger cell. The inequality of the size of the nuclei of the daugh-
ter cells, therefore, does not seem to be due to the unequal num-
ber of the chromosomes, but to an unequal quantity of cytoplasm.
In a case where the two daughter cells were about equal the size
of the nuclear membrane was about the same.

I have observed many cases in which the lagging chromosomes
appear to be divided, but I doubt that this ever occurs. Morgan
(1909) states: "That artificial conditions, such as handling or
osmosis, might break such a delicate connection at this time is not
at all improbable, and such an artificial result might give the im-
pression that the accessory is actually divided. Moreover, if the
bridge arches toward or away from the observer, the effect may
be produced at certain focal levels of discontinuity between the
ends of the lagging chromosomes, when none such exist."

The larger secondary spermatocyte receives eight divided and
two lagging undivided chromosomes, and the smaller secondary
spermatocyte receives eight divided chromosomes.

2. Larger Secondary Speniiatocyte.

The larger secondary spermatocyte undergoes an equal second
division without an intervening resting stage. The equatorial
plate (Figs. 12 and 13) of the second division shows ten chro-

352 H. HONDA.

mosomes. In the first maturation division five of the ten chromo-
somes are larger and five of them are smaller, but in this case six
are larger and four are smaller. The reason for this is discussed
later on. As in the case of the first division, chromosomes are
connected by linin threads. When the split chromosomes shift
to the opposite poles interzonal fibers appear. In the first divi-
sion the middle part of the two daughter cells becomes narrow
and show's an appearance of a dumb-bell with the ends different
in size. In this case, however, the middle part is broad, so that
the interzonal fibers are separated (Fig. 15).

3. Smaller Secondary Spermatocyte.

The smaller secondary spermatocyte shows chromosomes in its
nuclear cavity at the telophase of the first maturation division.
It is not difficult to distinguish the smaller secondary spermato-
cyte as their diameter is hardly half that of the larger ones. The
nucleus does not enter a resting stage. I have found in some
cases two small bodies near the nuclear membrane (Fig. 9).
These seem to be centrosomes, but I am unable to speak with cer-
tainty. The changes in preparation for the second division are
similar to those of the larger spermatocyte.

The equatorial plate (Figs. 20 and 21) shows eight chromo-
somes as compared with ten chromosomes in the equatorial plate
of the larger secondary spermatocyte. The cases which distinctly
show eight chromosomes are rare ; there can be little doubt, how-
ever, that this is the full number since there are ten chromosomes
in the equatorial plate of the first maturation division, and two
of them pass to the larger one as the lagging chromosomes. Four
chromosomes are larger and the other four are smaller. There
are five larger and five smaller chromosomes in the equatorial
plate of the first division ; the lagging chromosomes, therefore,
must be a larger and a smaller chromosome. If all the chromo-
somes were to divide in the first division, five larger and five
smaller chromosomes would appear in the equatorial plate of the
second division. Two chromosomes, one larger and one smaller,
lag and enter the larger cell without dividing. The smaller of
the lagging chromosomes, consequently, becomes larger than the

SPERM ATOGENESIS OF APHIDS. 353

other smaller chromosomes. This must be the reason why we
see four smaller chromosomes in the larger secondary spermato-
cyte instead of five.

The side view of the metaphase of the smaller spermatocyte
differs from that of the larger one in shape. It is more spindle-
shaped (Fig. 22). Fibers are not seen distinctly in the prepara-
tions stained with iron hematoxylin. The two stained bodies on
both sides of the chromosomes in the equatorial plate might be
the centrosomes (Fig. 23). There are cases which show sep-
arated chromosomes, and cases which show massed chromosomes
(Figs. 23-25). So far as my observation goes, in most cases the
chromosomes seem to fuse soon after their splitting. The telo-
phase does not show distinctly the interzonal fibers as in the case
of the larger cell. Two equal smaller spermatids are produced
after the division.

4. Smaller Spermatid.

The germ cells of each cyst of the testes are generally in about
the same stage. When the spermatids are young the cysts are
spherical in shape, but they elongate during the development of
the spermatids. The young smaller spermatids (Fig. 28) have
condensed nuclei, but the larger spermatids (Fig. 18) between
which they lie have vesicular nuclei. These smaller and larger
spermatids are seen all through the cyst. I have examined many
cases in order to see whether the polarities of the larger and
smaller spermatids are established with relation to the epithelium
or not. Most of the young larger and smaller spermatids, which
are seen near the epithelium, develop their tails toward the center
of the cyst, but some of them may develop along the epithelium
or develop their tails toward the epithelium. Those in the cen-
tral part do not show any definite orientation, and in extreme
cases spermatids existing side by side may show opposite direc-
tions. In cysts in which the larger spermatids are developed to
the stage shown in Fig. 18, the orientation of the larger and
smaller spermatids remains unchanged. In a little later stage,
however, all the larger and smaller spermatids begin to orient in
the same direction, and when the larger spermatids develop to
the stage shown in Fig. 19, all are oriented in the same direction.

354 H- HONDA.

There must be an interaction, probably chemical, between the
sustentacular cells and the larger and smaller spermatids. The
larger and smaller spermatids in the outer part opposite the sus-
tentacular cells and in the central part of the cyst generally move
among the tails of the other spermatids toward the sustentacular
cells, but those in the other parts move toward sustentacular cells
along the epithelium.

Developed smaller spermatids (Fig. 31) are seen among the
larger spermatids near the sustentacular cells, and do not show
any inferiority to the larger spermatids in moving toward the
cells. Before the nucleus of the larger spermatid showrs marked
differentiation the smaller spermatids have retreated a little
towards the interior. In other words, well developed smaller
spermatids approach towards the sustentacular cells, but do not
attach to them. I have examined many smaller spermatids in
order to see whether they develop apical parts. Figure 30 shows
a developing smaller spermatid, which has a cone-shaped apical
part. There is a developed smaller spermatid, which seems to
have a \vell-developed apical part, but we cannot distinctly ob-
serve since it is seen in close contact with the tails of the larger
spermatids. In most of the smaller spermatids, which have elon-
gated tails, I have not, however, observed developed apical parts.

As to the interpretation of the cells identified as smaller sper-
matids, may they not be degenerating larger spermatids? So
far as my observation goes the larger spermatids rarely degen-
erate ; moreover, it is not hard to distinguish degenerating young
larger spermatids from the smaller spermatids, since the former
are not only much larger than the latter, but the nucleus of the
larger spermatid becomes vesicular while the nucleus of the
smaller spermatid is condensed. If the larger spermatids devel-
oped to the stage shown in Fig. 19 begin to degenerate, we can
recognize them by the difference in the state of the nuclei. If the
almost fully developed spermatids begin to degenerate, it is quite
easy to tell them from the smaller spermatids, since they have
very slender nuclei and the smaller spermatids, which are seen in
the same cyst with them, have spherical nuclei. If degeneration
of the larger spermatids should occur at the stage in which they

SPERMATOGENESIS OF APHIDS. 355

I

have condensed ovoid nuclei which elongate later, the criterion
by which to distinguish them is their position, since when they
have developed to such a stage, the smaller spermatids with con-
densed spherical nuclei have already left the epithelium.

The metaphase of the smaller secondary spermatocytes are
seen among those of the larger secondary ones; I think, there-
fore, there is no doubt that the smaller secondary spermatocytes
undergo the second division. More developed larger spermatids
are seen with more developed smaller spermatids in the same
cyst. We may conclude from these observations that the smaller
spermatids develop with the larger spermatids.

I have observed cases where the larger and smaller sperma-
tids are seen in the central part of the cyst, while the majority of
spermatids have already reached the sustentacular cells. Such
larger and smaller spermatids might fail to reach the susten-
tacular cells, since they have to move among the spermatids. The
examination of the later stages, however, has shown that they
succeed in reaching the sustentacular cells.

Figure 32 shows a smaller spermatid which is abnormally big
and has a distinct axial filament. Ordinarily the smaller sperma-
tids elongate similarly, but are more delicate. One of the most
developed smaller spermatids is shown in Fig. 33. In such a
stage their development comes to a standstill, and they begin to
retrogress. They gradually retreat toward the tails of the larger
spermatids. Their nuclei which are deeply stained with iron
hematoxylin are seen among the tails of the larger spermatids
in a somewhat regular position. Finally they pass out to the
cavity -of the cyst.

The smaller spermatids fail in attaching to the sustentacular
cells ; they cannot, consequently, get material for their further
development. They have to live on their own substance. Their
tails become shorter, and the cytoplasm around the nucleus in-
creases (Fig. 34).

The forms shown in Fig. 35 are seen near the tail of the fully de-
veloped functional spermatozoa in the cavity of the cyst. We do
not see such spermatids in the cavities of the cysts at the younger
stages. These smaller spermatids still have elongated tails, but
later transform into spherical cells which have a distinct cell

356 H. HONDA.

membrane (Fig. 39) and show a tendency to fuse with each
other. There are some cells which have two or more condensed
nuclei. These seem to be the products of the fusion of two or
more smaller spermatids. Some of the retrogressed cells of the
smaller spermatids attach to the epithelium, and on these cells
other cells attach themselves ; thus they form layers as shown
in Fig. 38. In other cases they are irregularly attached to the
epithelium. When they attach themselves to each other they
show a polygonal shape.

A, b and c in Fig. 38 are parts of adjacent cysts, where fully
developed spermatozoa occur though not shown in the figure.
The cells occurring between the cysts are the retrogressed smaller
spermatids produced in the cyst c, and the epithelium proper is
very thin as seen between cysts a and b. As we see in the figure
these cells are not equal in size. In some of them the nuclei are
broken up and their fragments are seen scattered throughout the
cells. The others still show condensed spherical nuclei. As
stated already the larger spermatids rarely degenerate. These
larger spermatids may become like the cells just mentioned.
Though degenerating larger spermatids mingle among these cells,
there is no criterion by which they may be distinguished from
retrogressed cells of the smaller spermatids.

Some of these cells may be absorbed by the epithelial cells, but
how far the absorption proceeds is at present undetermined.
When these cells attach to the epithelium the functional sperma-
tozoa are already fully developed. Afterwards the wall of the
cyst ruptures, and these cells being deprived of their connection
with the testis are destined to disappear. It is possible that they
are extruded from the testis along with the spermatozoa. I have
observed epithelial cells of the cyst and retrogressed cells of the
smaller spermatids in some of the vasa deferentia. The sec-
tions of the testes of the old males show remarkable changes.
Their walls are thickened and neither spermatozoa nor the cysts,
which fill the young testes, can be seen.

IV. NEOTHOMASIA POPULICOLA AND MACROSIPHUM AMBROSIA.

The testes of embryos of Neothomasia populicola and Ma-cro-
siphum ambrosia are in the early stages, but those of larvae are

SPERMATOGENESIS OF APHIDS. 357

suitable for the purpose of studying the spermatocyte divisions.
As is the case in other aphid's, the primary spermatocytes of these
aphids divide unequally, and the anaphase shows the lagging chro-
mosomes. I have found in these aphids telophases of the second
maturation division of the smaller secondary spermatocyte, but
have observed no developing smaller spermatid ; we may, there-
fore, conclude that the smaller secondary spermatocytes and the
smaller spermatids of these aphids degenerate as in the cases of
the aphids studied by Morgan, von Baehr and Stevens. In Ma-
crosiphnni ambrosia I observed cases in which all smaller sec-
ondary spermatocytes seemed to be dividing, but I will conclude
in a succeeding paper whether all the smaller secondary sper-
matocytes divide or not.

In the cyst, where larger spermatids are already attached to the
sustentacular cells, there are seen spermatids which look like the
smaller spermatids of Stomaphis yanois. As stated above, since
no development of the smaller spermatids was observed, they
must be larger spermatids. In slightly younger cysts some sper-
matids are seen among the developed tails of other larger sper-
matids, which are about to attach to the sustentacular cells. Such
spermatids probably have no chance of reaching the cells. I have
found cases in which the larger spermatids are already attached
to the cells, but some spermatids are seen among the ends of the
tails of the larger spermatids. In other cysts spermatids with
condensed nuclei are seen apart from the sustentacular cells,
while others are attached to them.

As in the case of Stomaphis yanois young spermatids of these
aphids change their orientation to the same direction ; some sper-
matids, therefore, move to the sustentacular cells across the whole
diameter of the cyst or reach the cells moving along the epi-
thelium. If they move to the sustentacular cells along the epi-
thelium, as most of the spermatids do, they may lose the chance
to become attached to them. Developed spermatids have been
found by the side of spermatids which are attached to the sus-
tentacular cells and are developing. They wrere probably pre-
vented from reaching the cells by other spermatids, and their
development came to a standstill ; they, consequently, show
younger stages than the spermatids which are attached to the

358 H. HONDA.

cells. Since many spermatids are produced in the cysts, if they
move to the sustentacular cells through the tails of other sper-
matids, they meet much resistance; they, therefore, might be un-
able to reach the cells.

The most conspicuous difference between the case of Stomaphis
yanois and that of these aphids is the position of the retrogress-
ing spermatids. In the former case the smaller spermatids ap-
proach the sustentacular cells, and then gradually retreat toward
the tails of the larger spermatids ; their position, consequently, is
regular, having a relation to the development of the larger sper-
matids. In the latter case, however, the position of the retro-
gressing spermatids is irregular.

As in the case of Stomaphis yanois retrogressed spherical cells
are seen in the cyst with fully developed spermatozoa. These
cells attach themselves to the epithelium of the cysts and have the
same fate as the retrogressed cells of the smaller spermatids of
Stomaphis yanois.

V. REVIEW.

According to Meves and others, one of the secondary sperma-
tocytes of the honey bee is much smaller than the other, and re-
ceives no chromosomes ; it, consequently, degenerates after some
time. The larger secondary spermatocyte, moreover, divides
unequally in the second spermatocyte division. The chromo-
somes divide this time, and there are produced larger and smaller
spermatids. The larger spermatids differentiate into functional
spermatozoa. The smaller spermatids also undergo some differ-
entiation which, however, comes to a standstill at a late stage
and then they degenerate without transforming into functional
spermatozoa. The smaller spermatid of Stomaphis yanois re-
sembles that of the honey bee in some respects. Both of them are
much smaller than the larger spermatids, but judging from the
Meves' drawings, the difference in size between the larger and
the smaller spermatids is greater in the honey bee than in the
aphid. They both develop to some extent, but do not transform
into functional spermatozoa. Meves does not state what kind
of changes occurs in the degenerating smaller spermatids of the
honey bee ; I am, therefore, unable to compare their later stages
with those of the smaller spgrmatids of Stomaphis yanois.

SPERMATOGENESIS OF APHIDS. 359

The most conspicuous difference between the smaller sperma-
tids of the honey bee and this aphid is seen in the nuclei. The
nucleus of the smaller spermatid of the honey bee returns to a
resting stage, and differentiates similar to that of the larger
spermatid. The nucleus of the smaller spermatid of this aphid,
however, becomes condensed after the second spermatocyte divi-
sion, and remains in the same state, although the cytoplasm shows
changes similar to those of the larger spermatid. This may be
caused by the absence of the lagging chromosomes in the smaller
spermatids, while in the honey bee the smaller spermatids have
the same number of chromosomes as the larger spermatids.

Whitney (1918) mentions that the normal and rudimentary
spermatozoa have been found in considerable number of rotifers.
In his paper of 1917 he says that the functional spermatozoa are
identical in their power of determining the sex of the individual
that develops from a fertilized egg, since after a functional sper-
matozoon has fertilized a parthenogenic male egg, the egg always
develops into a female individual.

In the case of these rotifers, according to Whitney, the chro-
mosomes divide in the first spermatocyte division. One half of
the secondary spermatocytes divide and form the normal sper-
matids. The remaining half of the secondary spermatocytes,
contrary to the case of the smaller secondary spermatocyte of
Stomaphis yanois, do not divide, but develop directly into the
degenerate spermatozoa. The spermatocytes destined to degen-
erate are smaller than the others, and their development into the
complete rudimentary spermatozoa is strikingly different from
the development of the normal spermatids.

Whitney ('18) says that as all the fertilized eggs in both phyl-
loxerans and rotifers develop into female young, it seems safe to
conclude, as Morgan has already concluded, that the degenerate
sperm cells are the male-determining ones and that the normal
sperm cells are the female-determining ones.

Stevens (1905) found many degenerate spermatozoa in Blat-
tella germanlca. She states that the distribution and varying
number of these degenerate spermatozoa make it impossible to
interpret their condition as due to the absence of the accessory
chromosome as Miss Wallace does in the spider, and that the only

360 H. HONDA.

probable explanation seems to lie in the imperfect mitosis. She
detected no evidence of degeneracy among the young spermatids.

VI. SUMMARY.

1. In Stomaphis yanois the smaller secondary spermatocytes
divide, and develop to some extent, but retrogress to spherical
cells.

2. In Neothoniasia populicola and Macrosiphum ambrosia,
cases of division of the smaller secondary spermatocytes were
found, but no developing smaller spermatids were observed.

3. In Neothoniasia populicola and Macrosiphinn ambrosia
spherical cells like tho<se in Stomaphis yanois were found in the
cysts containing spermatozoa. These were identified as retro-
gressed larger spermatids.

LITERATURE.

von Baehr, W. B.

'08 Ueber die Bildung der Sexualzellen bei Aphididse. Zool. Anz., Bd. 33-

'09 Die Oogenese bei einigen viviparen Aphididen und die Spermatogenese
von Aphis saliceti, mit besonderer Berucksichtigung der chromatin-
verhaltnisse. Arch. f. Zellforsch., Bd. 3.

'12 Contribution a 1'etude de la caryocinese somatique de la pseudoreduc-

-tion. La Cellule, torn 27.
Meves, F.

'07 Die Spermatocytenteilungen bei der Honigbiene. Arch. f. mikr., Anat.

Bd. 70.
Morgan, T. H.

'08 The Production of Two Kinds of Spermatozoa in Phylloxerans —
functional "Female Producing" and Rudimentary Spermatozoa.
Proc. Soc. Exp. Biol. and Med., V.

'09 A Biological and Cytological Study of Sex Determination in Phyllox-
erans and Aphids. Jour. Exp. Zool., Vol. 7.

'10 The Chromosomes of the Parthenogenetic and Sexual Eggs of Phyl-
loxerans and Aphids. Proc. Soc. Exp. Biol. and Med., Vol. 7.

'15 The Determination of Sex in Phylloxerans and Aphids. Jour. Exp.

Zool., Vol. 19.
Stevens, N. M.

Study of Germ Cells of Aphis rosce and Aphis acnothera:. Jour. Exp.
Zool., Vol. 2.

Studies in Spennatogenesis with Special Reference to the " Acces-
sory Chromosome." Carnegie Inst. Wash., Pub. No. 36.

'o6a Studies of the Germ Cells of Aphids. Carnegie Inst. Wash., Pub.
No. 51.

'09 An Unpaired Heterochromosome in the Aphids. Jour. Exp. Zool..
Vol. 6.

SPERMATOGENESIS OF APHIDS. 361

Tannreuther, G. W.

'07 History of the Germ Cells and Early Embryology of Certain Aphids.

Zool. Jahrb. Anat. Ontog., Bd. 24.
Wallace, L. B.

'05 The Spermatogenesis of the Spider. BIOL. BULL., Vol. 6, No. 3.
Whitney, D. D.

'17 The Production of Functional and Rudimentary Spermatozoa in Roti-
fers. BIOL. BULL., Vol. 33.
'18 Further Studies on the Production of Functional and Rudimentary

Spermatozoa in Rotifers. BIOL. BULL., Vol. 34.
Wilson, H. V.

'n On the Behavior of Dissociated Cells in the Hydroids, Alcyonaria and
Asterias. Jour. Exp. Zool., Vol. n.

362 H. HONDA.

EXPLANATION OF PLATES.

All of the drawings were made with the aid of camera lucida. Figs, i
to 15 were drawn with a Leitz 1/16 oil immersion objective and a Zeiss com-
pensating ocular 18. Figs. 16 to 37. except Fig. 33, were drawn with a Leitz
1/16 oil immersion objective and a Leitz ocular 5. Fig. 33 was drawn with
a Leitz 1/16 oil immersion and a Leitz ocular 4. Fig. 38 was drawn with a
Leitz 1/16 oil immersion objective and a Leitz ocular 3. All figures from
Stonwphis yanois.

PLATE I.

FIGS, i AND 2. Primary spermatocytes, prophase.
FIGS. 3, 4, 5 AND 6. Primary spermatocytes, metaphase.
FIG. 7. Primary spermatocyte, anaphase.
FIGS. 8 AND 9. Primary spermatocytes, telophase.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE I.

H. HONDA.

364 H. HONDA.

PLATE II.

FIG. 10. Larger secondary spermatocytes, prophase.

FIGS. 11, 12 AND 13. Larger secondary spermatocyte, metaphase.

FIGS. 14 AND 15. Larger secondary spermatocytes, anaphase.

FIG. 16. Larger secondary spermatocyte, telophase.

FIGS. 17, 18 AND 19. Larger spermatids.

BIOLOGICAL BULLETIN, VOL. XL.

OLATt II.

13

18

H. HONDA.

366 H. HONDA.

PLATE III.

FIGS. 20, 21, 22, 23 AND 24. Smaller secondary spermatocytes, metaphase.
FIGS. 25 AND 26. Smaller secondary spermatocytes, anaphase.
FIG. 27. Smaller secondary spermatocyte, telophase.
FIGS. 28, 29, 30 AND 31. Smaller spermatids.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE III.

20

21

22

23

24

25

26

27

28

29

30

H. HONDA.

368 H. HONDA.

PLATE IV.

FIGS. 32 AND 33. Smaller spermatids.
FIGS. 34, 35 AND 36. Retrogressing smaller spermatids.
FIG. 37. Retrogressed cell of the smaller spermatid.
FIG. 38. Layers of the retrogressed cells.

BIOLOGICAL BULLETIN, VOL. XL.

PLATE IV.

33

34

35

36

37

38

H. HONDA.

M.?.L. WHOI LIBRARY

WH 17KC 3